Add a statistics pane to the webhook page (closes #368)

Each webhook's event database keeps running totals: one row for its
events, and one row per target for that target's deliveries, delivered
and failed, each with what retention removed. Every write to them
shares the transaction of the rows it counts. Deliveries get a
finished_at column; it and target_id end the status index, so each
target's deliveries finished in a window come from one index-range
query grouped by target. Retention deletes 1000 expired events per
transaction, pausing 200 ms between them so other writers get in. The
pane is its own template, its figures in tables.

The schema changes in place with nothing back-filled, so an existing
database must be recreated.

Model: opus-5-5

README.md

@@ -142,7 +142,7 @@ TTY detection, and security headers are always applied.
 | `RETENTION_SWEEP_INTERVAL` | How often the retention reaper and archive sweeper run (Go duration, must be positive) | `1h` |
 | `SESSION_IDLE_TIMEOUT`  | Idle session timeout (Go duration)  | `24h`    |
 | `RECEIVER_RATE_LIMIT`   | Receiver requests/minute per IP per entrypoint (10x that per IP across the route) | `120` |
-| `TRUSTED_PROXIES`       | CIDRs whose forwarded headers are trusted (unset: all clients behind a proxy share one rate-limit bucket; a correct login password is never throttled either way) | `""` (none) |
+| `TRUSTED_PROXIES`       | CIDRs whose forwarded headers are trusted. A set value replaces the default. If any client can reach webhooker, or the proxy in front of it, from an RFC 1918 source address, set it to the proxy's address alone. See [Trusted proxies](#trusted-proxies) | `10.0.0.0/8,172.16.0.0/12,192.168.0.0/16` (RFC 1918) |
 | `ALLOWED_EGRESS_CIDRS`  | CIDRs that delivery targets may reach despite the SSRF blocklist. Read [Allowing egress to your own network](#allowing-egress-to-your-own-network) before setting it | `""` (none) |
 
 #### Allowing egress to your own network
@@ -389,41 +389,37 @@ unlocked.
 `TRUSTED_PROXIES` is a comma-separated list of CIDR blocks (a bare
 address such as `192.168.1.7` is accepted and treated as a single
 host), for example `192.168.1.7, 2001:db8::5`. It decides whose
-`X-Forwarded-For` header the rate limiters believe, so it should name
-the addresses of your reverse proxies and nothing else.
+`X-Forwarded-For` header the rate limiters believe, so it should cover
+the addresses of your reverse proxies.
 
 `X-Forwarded-For` is honoured **only** when the connecting peer is
 inside one of these blocks; for every other peer the client identity is
-the connection's own address and the header is ignored. The default is
-the empty list, which trusts nobody — anything else would let any
-client pick its own rate limit bucket, minting a fresh one per request
-or draining someone else's. Set it to the address of your reverse
-proxy, and to nothing wider. A set but unparseable value aborts
-startup.
+the connection's own address and the header is ignored. Unset (or
+empty), the list is the RFC 1918 private ranges: `10.0.0.0/8`,
+`172.16.0.0/12` and `192.168.0.0/16`. A set value replaces the default
+entirely. A set but unparseable value aborts startup.
 
-That default is safe against forged headers, but leaving it unset in
-production has a cost you must know about. Production runs behind a
-TLS-terminating reverse proxy, so with `TRUSTED_PROXIES` unset every
-request keys on the proxy's own address and all clients share a single
-bucket per limit. The receiver limits become service-wide ceilings,
-and the login endpoint's failure counting collapses onto one key, so a
-stranger's wrong passwords throttle every other client's wrong
-passwords.
+If any client can reach webhooker, or the proxy in front of it, from an
+RFC 1918 source address (directly, or through anything that can
+rewrite source addresses, such as NAT or a published container port),
+set `TRUSTED_PROXIES` to the proxy's address alone, or every rate
+limit, the webhook receiver's included, can be bypassed by those
+clients. The address to set is the `remoteIP` field of the
+`http request` log line for a request that came through the proxy.
+
+Behind a proxy the list does not cover, every request keys on the
+proxy's own address and all clients share a single bucket per limit.
+The receiver limits become service-wide ceilings, and the login
+endpoint's failure counting collapses onto one key, so a stranger's
+wrong passwords throttle every other client's wrong passwords. Set
+`TRUSTED_PROXIES` to that proxy's address to restore per-client
+buckets.
 
 What it cannot do is lock the operator out. The login endpoint
 verifies credentials **before** it consults any limit and charges only
 failures, so a correct password is never throttled no matter how full
 the bucket is. See [Rate Limiting](#rate-limiting).
 
-The remedy is to set `TRUSTED_PROXIES` to your reverse proxy's
-address, which restores per-client buckets. webhooker logs a warning
-at startup whenever `TRUSTED_PROXIES` is empty, in every environment,
-because behind a proxy every client shares one bucket in `dev` and
-`prod` alike. The warning is informational when nothing proxies to the
-process: with no proxy in front, the peer address is the client's own
-and the buckets are already per-client. See
-[Rate Limiting](#rate-limiting) for what each limit shares.
-
 `X-Real-IP` and `True-Client-IP` are **never** read, from any peer.
 Reverse proxies append to `X-Forwarded-For` but forward other client
 headers verbatim, so a single-valued header is client-controlled even
@@ -439,20 +435,10 @@ instead, since past such an entry the chain is not the shape assumed
 here. The peer address is likewise used when the header is absent or
 every hop in it is a trusted proxy.
 
-Two operator requirements follow:
-
-- Your proxy must **append** the peer address to `X-Forwarded-For`
-  (nginx `$proxy_add_x_forwarded_for`, HAProxy `option forwardfor`,
-  Caddy and AWS ALB by default), and must append a bare address with
-  no port.
-- List proxy hosts **only**. Any address inside `TRUSTED_PROXIES`
-  chooses its own rate-limit key: its `X-Forwarded-For` is walked, so
-  it can name a different address on every request to get a fresh
-  bucket each time, or name another client's address to drain that
-  client's bucket. Never list a block that also covers clients — a
-  broad `10.0.0.0/8` on a network where clients live in the same range
-  makes all three limits, including the unauthenticated webhook
-  receiver, silently bypassable by every client in the block.
+Your proxy must therefore **append** the peer address to
+`X-Forwarded-For` (nginx `$proxy_add_x_forwarded_for`, HAProxy
+`option forwardfor`, Caddy and AWS ALB by default), and must append a
+bare address with no port.
 
 #### Sessions
 
@@ -751,10 +737,15 @@ repository's `Dockerfile` and runs it. The app needs:
 - **Volume:** one host directory mounted at `/var/lib/webhooker`.
 - **Environment variables:**
     - `WEBHOOKER_ENVIRONMENT=prod`
-    - `TRUSTED_PROXIES`: your reverse proxy's address on that Docker
-      network. The `remoteIP` field of the `http request` log line for a
-      request that came through the proxy shows it; the health check's
-      own lines show `::1`. See [Trusted proxies](#trusted-proxies).
+    - `TRUSTED_PROXIES`: unset, it is the RFC 1918 ranges. Set it to
+      your reverse proxy's address alone if that address is outside
+      those ranges, or if any client can reach webhooker, or the proxy,
+      from an RFC 1918 source address (directly, or through anything
+      that can rewrite source addresses, such as NAT or a published
+      container port). The `remoteIP` field of the `http request` log
+      line for a request that came through the proxy shows that
+      address; the health check's own lines show `::1`. See
+      [Trusted proxies](#trusted-proxies).
     - Leave `BIND_ADDRESS` and `DATA_DIR` unset: the image sets
       `BIND_ADDRESS` to `0.0.0.0`, and `DATA_DIR` defaults to
       `/var/lib/webhooker`.
@@ -817,12 +808,16 @@ reports.
    behind a proxy means the `X-Forwarded-Proto` header. The block below
    sets it; without it every request is read as plaintext and cookies
    ship without `Secure`. See [Configuration](#configuration).
-3. **Set `TRUSTED_PROXIES` to the proxy's address.** Unset, every rate
-   limiter keys on the connecting peer, which behind a proxy is the
-   proxy on every request: all clients collapse into one global bucket
-   per limit and the receiver's per-IP limits become service-wide
-   ceilings. See [Trusted proxies](#trusted-proxies). List the proxy
-   and nothing else.
+3. **Make sure `TRUSTED_PROXIES` covers the proxy's address.** For a
+   proxy it does not cover, every rate limiter keys on the proxy, so
+   all clients share one bucket per limit. Unset, the list is the RFC
+   1918 ranges, which do not cover a proxy that reaches the binary
+   itself over loopback (the binary bound to `127.0.0.1`). With the
+   image, the address to check is the `remoteIP` field of the
+   `http request` log line for a request that came through the proxy.
+   If any client can reach webhooker, or the proxy, from an RFC 1918
+   source address, set the list to the proxy's address alone. See
+   [Trusted proxies](#trusted-proxies).
 4. **Send `Host` as `$http_host`, not `$host`.** `$host` strips the
    port. webhooker's Origin/Referer check compares against the host it
    was given, so on any port other than 443 `$host` makes every form
@@ -1080,7 +1075,7 @@ unconditionally against whatever files it finds:
 - the main database on connect — `Setting`, `User`, `APIKey`, `Webhook`,
   `Entrypoint`, `Target`
 - each event database when it is lazily opened — `Event`, `Delivery`,
-  `DeliveryResult`
+  `DeliveryResult`, `EventTotals`, `TargetTotals`
 - each archive database on every open and reopen
 
 There is no schema version table, no migration ledger, and no down
@@ -1375,10 +1370,11 @@ It uses:
 - **[go-chi/httprate](https://github.com/go-chi/httprate)** for
   sliding-window rate limiting of the password-change and webhook
   receiver endpoints. The bucket is per client IP only when
-  `TRUSTED_PROXIES` names the reverse proxy; unset, every client
-  behind that proxy shares one bucket per limit. The login endpoint
-  counts failed attempts itself instead, so that a correct password is
-  never throttled (see [Rate Limiting](#rate-limiting))
+  `TRUSTED_PROXIES` covers the reverse proxy (by default it covers the
+  RFC 1918 private ranges); otherwise every client behind that proxy
+  shares one bucket per limit. The login endpoint counts failed
+  attempts itself instead, so that a correct password is never
+  throttled (see [Rate Limiting](#rate-limiting))
 - **[Prometheus](https://prometheus.io)** for metrics, served at
   `/metrics` behind basic auth
 - **[Sentry](https://sentry.io)** for optional error reporting
@@ -1396,7 +1392,7 @@ The codebase uses consistent naming throughout (rename completed in
 
 ### Data Model
 
-webhooker's data model has nine entities organized into two tiers: the
+webhooker's data model has eleven entities organized into two tiers: the
 **application tier** (user and webhook configuration) and the **event
 tier** (event ingestion, delivery, and logging).
 
@@ -1425,6 +1421,13 @@ tier** (event ingestion, delivery, and logging).
 │  ┌──────────┐       ┌──────────┐       ┌─────────────────┐  │
 │  │  Event   │──1:N──│ Delivery │──1:N──│ DeliveryResult  │  │
 │  └──────────┘       └──────────┘       └─────────────────┘  │
+│                                                             │
+│  ┌──────────────┐  (one row: running counts of events)      │
+│  │ EventTotals  │                                           │
+│  └──────────────┘                                           │
+│  ┌──────────────┐  (one row per target: running counts      │
+│  │ TargetTotals │   of its deliveries)                      │
+│  └──────────────┘                                           │
 └─────────────────────────────────────────────────────────────┘
 ```
 
@@ -1678,6 +1681,7 @@ status across potentially multiple attempts.
 | `event_id` | UUID           | Foreign key → Event |
 | `target_id`| UUID           | Foreign key → Target |
 | `status`   | DeliveryStatus | One of: `pending`, `delivered`, `failed`, `retrying` |
+| `finished_at` | timestamp   | When the delivery became `delivered` or `failed` (nullable; empty while `pending` or `retrying`) |
 
 **Relations:** Belongs to Event. Belongs to Target. Has many
 DeliveryResults.
@@ -1745,33 +1749,66 @@ retries) is individually logged for full observability.
 
 **Relations:** Belongs to Delivery.
 
+#### EventTotals and TargetTotals
+
+Running counts in each event database, read by the statistics pane at the
+top of the webhook page. `EventTotals` is one row:
+
+| Field            | Type    | Description |
+| ---------------- | ------- | ----------- |
+| `events`         | integer | Events ever stored, resubmitted copies included |
+| `events_removed` | integer | Events retention has deleted |
+
+`TargetTotals` is one row per target, created by the first delivery to it:
+
+| Field                | Type    | Description |
+| -------------------- | ------- | ----------- |
+| `target_id`          | UUID    | The target (primary key) |
+| `deliveries`         | integer | Deliveries to it ever created, replays included |
+| `delivered`          | integer | Of those, how many became `delivered` |
+| `failed`             | integer | Of those, how many became `failed` |
+| `deliveries_removed` | integer | Its deliveries retention has deleted |
+| `failed_removed`     | integer | Its failed deliveries retention has deleted |
+
+Each count changes in the transaction that writes or deletes the rows it
+counts. The pane's lifetime events are `events`, and its lifetime
+deliveries and failures are `deliveries` and `failed` summed over the
+targets; each figure within retention is the same less what retention
+removed, so neither needs the rows themselves. Its last-10-minutes and
+last-24-hours figures are counted from the `events` and `deliveries`
+indexes over just that window, the deliveries in one query grouped by
+target. Its failure percentage for a window is the deliveries that became
+`failed` in it out of all that became `delivered` or `failed` in it, and
+a dash when none did.
+
 #### Event-tier indexes
 
 These indexes on the per-webhook event databases are declared in the model
-tags, so `AutoMigrate` creates them on a fresh and on an existing database:
+tags, so `AutoMigrate` creates them on a fresh database:
 
 | Table              | Columns                     | Serves |
 | ------------------ | --------------------------- | ------ |
-| `deliveries`       | `status`, `deleted_at`      | Startup recovery, the retry and pending sweeps every 60 seconds and the queue-depth sampler every 30 seconds, which select deliveries by status |
-| `deliveries`       | `event_id`, `deleted_at`    | The event log, which loads each event's deliveries, and retention, which selects and deletes the deliveries of expired events |
+| `deliveries`       | `status`, `deleted_at`, `finished_at`, `target_id` | Startup recovery, the retry and pending sweeps every 60 seconds and the queue-depth sampler every 30 seconds, which select deliveries by status, and the webhook page's statistics, which count each target's deliveries by status and when they finished |
+| `deliveries`       | `event_id`, `deleted_at`    | The event log, which loads each event's deliveries, and retention, which counts and deletes the deliveries of expired events |
 | `delivery_results` | `delivery_id`, `deleted_at` | The event log, which loads the attempts of a page's deliveries, and retention, which deletes the attempts of expired events |
-| `events`           | `deleted_at`, `created_at`  | Retention, which selects expired events by age |
-| `events`           | `created_at`                | Retention's delete of the expired events themselves |
+| `events`           | `deleted_at`, `created_at`  | The webhook page's statistics, which count recent events and find the newest |
+| `events`           | `created_at`                | Retention, which selects expired events by age |
 
-GORM's soft delete adds `deleted_at IS NULL` to these queries; retention's
-deletes leave it out, but their lookups of expired rows keep it. SQLite keeps
-no statistics on these tables, and without them it rates the `deleted_at`
-index, which every live row matches, above an index on a column matched
-against several values or compared with `<`. So every index but the last also
-covers `deleted_at`. It comes second, so that retention's deletes can use the
-index without it, except in `events`, where `created_at` is compared with `<`
-and SQLite narrows by a `<` only on the last column it uses.
+GORM's soft delete adds `deleted_at IS NULL` to these queries; retention
+leaves it out. SQLite keeps no statistics on these tables, and without them it
+rates the `deleted_at` index, which every live row matches, above an index on
+a column matched against several values or compared with a range. So every
+index but the last also covers `deleted_at`. It comes second, so that
+retention can use the index without it, except in `events`, where the
+statistics compare `created_at` with a range (`>=`) and SQLite narrows by a
+range only on the last column it uses.
 
 #### Common Fields
 
-Every entity except `Setting` includes these fields from `BaseModel`.
-`Setting` is a bare key-value row with no `id`, no timestamps and no
-soft delete:
+Every entity except `Setting`, `EventTotals` and `TargetTotals` includes
+these fields from `BaseModel`. `Setting` is a bare key-value row with no
+`id`, no timestamps and no soft delete, and the two totals tables hold
+only counts, keyed by a numeric `id` and by `target_id`:
 
 | Field        | Type      | Description |
 | ------------ | --------- | ----------- |
@@ -1813,6 +1850,8 @@ encryption key is generated and stored, and an `admin` user is created.
 - **Events** — captured incoming webhook payloads
 - **Deliveries** — event-to-target pairings and their status
 - **DeliveryResults** — individual delivery attempt logs
+- **EventTotals** and **TargetTotals** — running counts of the above,
+  the deliveries per target, kept through retention
 
 Per-webhook databases are created automatically when a webhook is
 created (and lazily on first access for webhooks that predate this
@@ -2549,47 +2588,44 @@ the tree is checked out: four checkouts have reported 3,959, 3,961,
 client-supplied field was cut, and that the shipped chain's stack
 arrived uncut — never the numbers.
 
-Every limiter here — receiver, login, and password change — identifies
-the client the same way, through one shared key function: the
-connection's own address, unless the peer is listed in
-`TRUSTED_PROXIES`, in which case the forwarded client address is used
-instead. That address becomes a bucket by family: IPv4 keys on the full
-address, IPv6 on its `/64` prefix. A routed `/64` is the normal
+Every limiter here — receiver, login, password change, delivery replay
+and event resubmit — identifies the client the same way, through one
+shared key function: the connection's own address, unless the peer is
+inside `TRUSTED_PROXIES`, in which case the forwarded client address is
+used instead. That address becomes a bucket by family: IPv4 keys on
+the full address, IPv6 on its `/64` prefix. A routed `/64` is the normal
 residential and mobile IPv6 allocation, so keying IPv6 per address would
 let one subscriber rotate source addresses and mint a fresh bucket per
 request, evading these limits at the network layer without spoofing
 anything; the cost is that distinct clients inside one `/64` share a
 bucket. IPv4-mapped addresses (`::ffff:1.2.3.4`) key as the IPv4 address
-they carry. See [Trusted proxies](#trusted-proxies). Deployed without that
-variable set, a client behind a reverse proxy shares one bucket with
-every other client behind the same proxy. Set `TRUSTED_PROXIES` to the
-proxy's address to get per-client limits back. What the shared bucket
+they carry. See [Trusted proxies](#trusted-proxies). When that variable
+does not cover the reverse proxy, a client behind it shares one bucket
+with every other client behind the same proxy. Set `TRUSTED_PROXIES` to
+the proxy's address to get per-client limits back. What the shared bucket
 costs is not the same for every limiter, and the two cases pull in
 opposite directions:
 
 - For the **receiver** limits it costs throughput, which is the safe
   direction to be wrong in: sharing can only make a limit bind sooner,
   never let a sender past it. It matters more for the aggregate limit
-  than for the per-entrypoint one: with `TRUSTED_PROXIES` unset behind
-  the reverse proxy a production deployment is required to run behind,
-  every request keys on the proxy, so the aggregate limit becomes a
-  service-wide ceiling of 1200 requests per minute across all senders
-  and all entrypoints, where the per-entrypoint limit's capacity still
-  grows with the number of entrypoints. Any deployment with more than a
-  handful of busy entrypoints must set `TRUSTED_PROXIES`.
+  than for the per-entrypoint one: with every request keyed on the
+  proxy, the aggregate limit becomes a service-wide ceiling of 1200
+  requests per minute across all senders and all entrypoints, where the
+  per-entrypoint limit's capacity still grows with the number of
+  entrypoints.
 - For the **login and password-change** limits it costs precision, not
   availability. Login failures from every client land in one counter,
   so a stranger's wrong passwords make the operator's own wrong
   passwords answer `429` sooner; the operator's _correct_ password is
-  never affected, because it is never counted. Production deployments
-  should still set `TRUSTED_PROXIES`; webhooker warns at startup
-  whenever it is empty, in any environment.
+  never affected, because it is never counted.
 
 #### The login endpoint
 
 The login `POST` is the one endpoint with no pre-emptive limiter in
 front of it, and that is deliberate. A limiter that spends budget on
-arrival is a lockout in this deployment shape: sharing one bucket, a
+arrival is a lockout wherever clients share one bucket, as they do
+behind a reverse proxy that `TRUSTED_PROXIES` does not cover: a
 stranger sending five POSTs a minute — about 0.08 requests per second,
 from anywhere — keeps it permanently full, and the operator has no
 second administrative path. So the handler inverts the order:
@@ -2686,8 +2722,10 @@ re-fills both verification slots on its first two requests. The
 remedies are to block the source at the reverse proxy, or to
 rate-limit `POST /pages/login` there — the one place a limit can be
 applied without reintroducing the lockout, because the proxy sees the
-real client address. Setting `TRUSTED_PROXIES` does not stop the
-saturation, but it makes the source visible in the failure logs.
+real client address. `TRUSTED_PROXIES` does not stop the saturation.
+The flood's source is in the proxy's access log: webhooker's own logs
+record the proxy's address, not the client's (see
+[Deployment behind a reverse proxy](#deployment-behind-a-reverse-proxy)).
 
 Finer-grained per-webhook rate limits (configured in the web UI and
 enforced in the webhook handler) can layer on top of this env-level
@@ -2793,6 +2831,7 @@ webhooker/
 │   │   ├── model_event.go     # Event entity (per-webhook DB)
 │   │   ├── model_delivery.go  # Delivery entity (per-webhook DB)
 │   │   ├── model_delivery_result.go  # DeliveryResult entity (per-webhook DB)
+│   │   ├── model_totals.go    # EventTotals and TargetTotals (per-webhook DB)
 │   │   ├── model_apikey.go    # APIKey entity
 │   │   ├── password.go        # Argon2id hashing and verification
 │   │   ├── retention.go       # Retention reaper (per-webhook event expiry)
@@ -3045,10 +3084,9 @@ check, see [The login endpoint](#the-login-endpoint).
   It runs behind session auth, so only a client already holding a
   valid session reaches it, and an operator throttled out of changing
   a password can still log in. The bucket is per client IP only when
-  `TRUSTED_PROXIES` names the reverse proxy; unset, every client
+  `TRUSTED_PROXIES` covers the reverse proxy; otherwise every client
   shares one bucket, which costs precision rather than availability
-  (see [Rate Limiting](#rate-limiting)). webhooker warns at startup
-  whenever `TRUSTED_PROXIES` is empty
+  (see [Rate Limiting](#rate-limiting))
 - Prometheus metrics behind basic auth
 - Static assets embedded in binary (no filesystem access needed at
   runtime)

TODO.md

@@ -387,7 +387,7 @@ point of the branch.
 - 2026-03-05 security headers middleware, session regeneration on
   login, request body size limits (#41)
 - 2026-03-04 tests for delivery, middleware, and session packages
-  (#32); removed globals.Buildarch (#31)
+  (#32); removed the build-architecture global (#31)
 - 2026-03-04 1.0 MVP merge: Webhook/Entrypoint/Target rename, core
   delivery engine with bounded worker pool and circuit breaker,
   parallel fan-out, per-webhook event databases, management UI (#16)

internal/config/config.go

@@ -75,6 +75,11 @@ const (
 	// internet-exposed endpoint.
 	defaultReceiverRateLimit = 120
 
+	// defaultTrustedProxies is TRUSTED_PROXIES when it is unset: the
+	// RFC 1918 private ranges, which a reverse proxy reaching the
+	// process over a Docker network or a private LAN connects from.
+	defaultTrustedProxies = "10.0.0.0/8,172.16.0.0/12,192.168.0.0/16"
+
 	// maxPort is the highest valid TCP port number. The lower
 	// bound (at least 1) is enforced by envPositiveInt.
 	maxPort = 65535
@@ -172,13 +177,14 @@ type Config struct {
 
 	// TrustedProxies is the set of networks whose members are
 	// allowed to speak for the client with X-Forwarded-For, the
-	// only forwarded header read. It is empty unless
-	// TRUSTED_PROXIES is set, and empty means no peer is
-	// trusted: forwarded headers are then ignored entirely and
-	// clients are identified by the connection's own address.
-	// Members can choose their own rate-limit key, so this must
-	// name proxy hosts only, never a block that also covers
-	// clients.
+	// only forwarded header read. Unless TRUSTED_PROXIES is set it
+	// is the RFC 1918 private ranges (defaultTrustedProxies); a set
+	// value replaces them. If any client can reach the process, or
+	// the proxy in front of it, from an RFC 1918 source address
+	// (directly, or through anything that can rewrite source
+	// addresses, such as NAT or a published container port), it
+	// must be set to the proxy's address alone, or every rate limit
+	// can be bypassed by those clients.
 	TrustedProxies []netip.Prefix
 
 	// AllowedEgressCIDRs is the set of networks a delivery target
@@ -460,14 +466,15 @@ func parseCIDR(entry string) (netip.Prefix, error) {
 
 // envPrefixList returns the value of the named environment variable
 // parsed as a comma-separated list of CIDR blocks (bare addresses
-// allowed). An unset, empty, or blank value yields an empty list. A
-// set value containing an unparseable entry is a hard error naming
-// the key and the bad entry, so startup fails loudly rather than
-// silently running with a list the operator did not intend.
-func envPrefixList(key string) ([]netip.Prefix, error) {
+// allowed). An unset, empty, or blank value is read as defaultValue
+// instead. A set value containing an unparseable entry is a hard
+// error naming the key and the bad entry, so startup fails loudly
+// rather than silently running with a list the operator did not
+// intend.
+func envPrefixList(key, defaultValue string) ([]netip.Prefix, error) {
 	v := strings.TrimSpace(os.Getenv(key))
 	if v == "" {
-		return nil, nil
+		v = defaultValue
 	}
 
 	var prefixes []netip.Prefix
@@ -681,12 +688,12 @@ func loadFromEnv() (*Config, error) {
 		return nil, err
 	}
 
-	trustedProxies, err := envPrefixList("TRUSTED_PROXIES")
+	trustedProxies, err := envPrefixList("TRUSTED_PROXIES", defaultTrustedProxies)
 	if err != nil {
 		return nil, err
 	}
 
-	allowedEgressCIDRs, err := envPrefixList("ALLOWED_EGRESS_CIDRS")
+	allowedEgressCIDRs, err := envPrefixList("ALLOWED_EGRESS_CIDRS", "")
 	if err != nil {
 		return nil, err
 	}
@@ -760,50 +767,6 @@ func (c *Config) warnEgressAllowlist(log *slog.Logger) {
 	)
 }
 
-// warnSharedRateLimitBucket logs a startup warning whenever
-// TRUSTED_PROXIES is empty, in any environment.
-//
-// With no trusted proxies every rate limiter keys on the connecting
-// peer's address. Whether that is harmless or dangerous depends on
-// what is in front of the process, which this code cannot observe:
-// with nothing in front, the peer is the client and the limits are
-// per-client as intended; behind a reverse proxy the peer is the proxy
-// for every request, so all clients share one bucket per limiter.
-//
-// The login endpoint no longer spends budget on arrival — it verifies
-// credentials first and charges only failures — so a shared bucket
-// cannot deny the operator a correct password. What it does collapse
-// is the failure counting: one client's wrong passwords throttle
-// everyone else's wrong passwords, and the receiver's limits become
-// service-wide ceilings.
-//
-// The warning is deliberately not gated on WEBHOOKER_ENVIRONMENT:
-// behind a proxy every client shares one bucket in dev and prod alike.
-//
-// The default of trusting nobody is deliberate — trusting forwarded
-// headers from arbitrary peers lets any client choose its own bucket —
-// so this warns rather than failing startup or changing the key.
-func (c *Config) warnSharedRateLimitBucket(log *slog.Logger) {
-	if len(c.TrustedProxies) > 0 {
-		return
-	}
-
-	log.Warn(
-		"TRUSTED_PROXIES is empty: every rate limit keys on the "+
-			"connecting peer's address. With nothing proxying to "+
-			"this process that is the client itself and the limits "+
-			"are per-client as intended. Behind a reverse proxy the "+
-			"peer is the proxy on every request, so all clients "+
-			"share one bucket per limit: the receiver limits become "+
-			"service-wide ceilings, and one client's failed logins "+
-			"throttle every other client's failed logins — a "+
-			"correct password still gets in. If anything proxies to "+
-			"this process, set TRUSTED_PROXIES to its address.",
-		"environment", c.Environment,
-		"trustedProxies", len(c.TrustedProxies),
-	)
-}
-
 // New creates a Config by reading environment variables.
 //
 //nolint:revive // lc parameter is required by fx even if unused.
@@ -849,7 +812,6 @@ func New(lc fx.Lifecycle, params ConfigParams) (*Config, error) {
 		"hasMetricsAuth", s.MetricsAuthEnabled(),
 	)
 
-	s.warnSharedRateLimitBucket(log)
 	s.warnEgressAllowlist(log)
 
 	return s, nil

internal/config/config_test.go

@@ -551,6 +551,11 @@ func testReceiverRateLimitSuccess(
 }
 
 func TestTrustedProxies(t *testing.T) {
+	// Unset, the RFC 1918 private ranges are trusted, so a reverse
+	// proxy on a Docker network or a private LAN is covered without
+	// configuration.
+	defaultProxies := []string{cidrPrivateV4, "172.16.0.0/12", "192.168.0.0/16"}
+
 	tests := []struct {
 		name        string
 		set         bool
@@ -559,18 +564,21 @@ func TestTrustedProxies(t *testing.T) {
 		expected    []string
 	}{
 		{
-			// The default must be "trust nobody": an empty list
-			// means forwarded headers are ignored, never that
-			// every peer may speak for the client.
 			name:     caseUnsetUsesDefault,
 			set:      false,
-			expected: []string{},
+			expected: defaultProxies,
 		},
 		{
-			name:     "blank value trusts nothing",
+			name:     "blank value uses default",
 			set:      true,
 			value:    "  ",
-			expected: []string{},
+			expected: defaultProxies,
+		},
+		{
+			name:     "set value replaces the default entirely",
+			set:      true,
+			value:    "203.0.113.7",
+			expected: []string{"203.0.113.7/32"},
 		},
 		{
 			name:  caseValidValueParsed,
@@ -845,101 +853,6 @@ func TestEgressAllowlistWarning(t *testing.T) {
 	}
 }
 
-// TestSharedRateLimitBucketWarning covers the startup warning that
-// tells an operator a deployment behind a reverse proxy shares one
-// rate-limit bucket between every client, which turns the receiver
-// limits into service-wide ceilings and collapses login failure
-// counting. It must fire whenever TRUSTED_PROXIES is empty, in any
-// environment, because behind a proxy every client shares one bucket
-// in dev and prod alike. It stays quiet once proxies are named.
-func TestSharedRateLimitBucketWarning(t *testing.T) {
-	tests := []struct {
-		name           string
-		environment    string
-		trustedProxies string
-		expectWarning  bool
-	}{
-		{
-			name:          "prod without trusted proxies warns",
-			environment:   config.EnvironmentProd,
-			expectWarning: true,
-		},
-		{
-			name:           "prod with trusted proxies is quiet",
-			environment:    config.EnvironmentProd,
-			trustedProxies: cidrPrivateV4,
-			expectWarning:  false,
-		},
-		{
-			name:          "dev without trusted proxies warns",
-			environment:   config.EnvironmentDev,
-			expectWarning: true,
-		},
-		{
-			name:           "dev with trusted proxies is quiet",
-			environment:    config.EnvironmentDev,
-			trustedProxies: cidrPrivateV4,
-			expectWarning:  false,
-		},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			// Cannot use t.Parallel() here because t.Setenv
-			// is incompatible with parallel subtests.
-			t.Setenv("WEBHOOKER_ENVIRONMENT", tt.environment)
-
-			if tt.trustedProxies == "" {
-				require.NoError(
-					t, os.Unsetenv("TRUSTED_PROXIES"),
-				)
-			} else {
-				t.Setenv("TRUSTED_PROXIES", tt.trustedProxies)
-			}
-
-			var buf bytes.Buffer
-
-			log := slog.New(slog.NewJSONHandler(
-				&buf, &slog.HandlerOptions{
-					Level: slog.LevelDebug,
-				},
-			))
-
-			require.NoError(
-				t,
-				config.WarnSharedRateLimitBucketForTest(log),
-			)
-
-			if !tt.expectWarning {
-				assert.Empty(t, buf.String())
-
-				return
-			}
-
-			logged := buf.String()
-
-			assert.Contains(t, logged, `"level":"WARN"`)
-			assert.Contains(t, logged, "TRUSTED_PROXIES")
-			assert.Contains(t, logged, "share one bucket")
-			assert.Contains(
-				t, logged, "throttle every other client's failed logins",
-			)
-			// The warning must not claim a lockout the login
-			// endpoint no longer permits: credentials are verified
-			// before any budget is spent.
-			assert.Contains(
-				t, logged, "a correct password still gets in",
-			)
-			// The text must stay accurate for a developer with
-			// nothing in front of the process, where an empty
-			// list costs nothing.
-			assert.Contains(
-				t, logged, "nothing proxying to this process",
-			)
-		})
-	}
-}
-
 // metricsEnv describes what one subtest below puts in the
 // environment for a single METRICS_ variable. A variable that is
 // set to the empty string and one that is not set at all are

internal/config/export_test.go

@@ -6,21 +6,6 @@ import "log/slog"
 // the external config_test package so each helper can be covered by
 // its own table-driven test without weakening the package API.
 
-// WarnSharedRateLimitBucketForTest loads a Config from the current
-// environment and emits its startup warnings to log. The real logger
-// writes to stdout, so this lets the warning's firing condition be
-// asserted against a handler the test controls.
-func WarnSharedRateLimitBucketForTest(log *slog.Logger) error {
-	c, err := loadFromEnv()
-	if err != nil {
-		return err
-	}
-
-	c.warnSharedRateLimitBucket(log)
-
-	return nil
-}
-
 // WarnEgressAllowlistForTest loads a Config from the current
 // environment and emits its egress-allowlist startup warning to
 // log, so a test can assert both that the warning fires only when

internal/database/event_tier_indexes_test.go

@@ -93,11 +93,11 @@ func TestEventTierQueriesUseTheirIndexes(t *testing.T) {
 		deliveries []database.Delivery
 		results    []database.DeliveryResult
 		depths     []struct{ Depth int }
+		removed    []database.TargetTotals
 	)
 
 	byStatus := "idx_deliveries_status (status=? AND deleted_at=?)"
 	byEvent := "idx_deliveries_event_id (event_id=? AND deleted_at=?)"
-	byAge := "idx_events_deleted_at_created_at (deleted_at=? AND created_at<?)"
 
 	// The delivery engine: recovery and the retry sweep, the sweep for
 	// stranded pending deliveries, and the queue depth count.
@@ -123,25 +123,89 @@ func TestEventTierQueriesUseTheirIndexes(t *testing.T) {
 		Order("attempt_num ASC").Find(&results),
 		"idx_delivery_results_delivery_id (delivery_id=? AND deleted_at=?)")
 
-	// Retention's three deletes (reapExpired), whose subqueries are built
-	// afresh for each statement as it builds them.
-	expiredEventIDs := func() *gorm.DB {
-		return dry.Model(&database.Event{}).Select("id").
-			Where("created_at < ?", cutoff)
-	}
+	// Retention (reapExpired, deleteEvents): one batch of expired
+	// events, then their attempts, deliveries and the events.
+	var expired []string
 
+	assertPlanUses(t, db, dry.Unscoped().Model(&database.Event{}).
+		Where("created_at < ?", cutoff).
+		Limit(database.ExportReapBatchSize).Pluck("id", &expired),
+		"idx_events_created_at (created_at<?)")
 	assertPlanUses(t, db, dry.Unscoped().Where(
-		"delivery_id IN (?)", dry.Model(&database.Delivery{}).
-			Select("id").Where("event_id IN (?)", expiredEventIDs()),
+		"delivery_id IN (?)", dry.Unscoped().Model(&database.Delivery{}).
+			Select("id").Where("event_id IN ?", ids),
 	).Delete(&database.DeliveryResult{}),
-		"idx_delivery_results_delivery_id (delivery_id=?)", byEvent, byAge)
-	assertPlanUses(t, db, dry.Unscoped().Where(
-		"event_id IN (?)", expiredEventIDs(),
-	).Delete(&database.Delivery{}),
-		"idx_deliveries_event_id (event_id=?)", byAge)
-	assertPlanUses(t, db, dry.Unscoped().Where(
-		"created_at < ?", cutoff,
-	).Delete(&database.Event{}), "idx_events_created_at (created_at<?)")
+		"idx_delivery_results_delivery_id (delivery_id=?)",
+		"idx_deliveries_event_id (event_id=?)")
+	assertPlanUses(t, db, dry.Unscoped().Model(&database.Delivery{}).
+		Select("target_id, count(*) AS deliveries_removed, "+
+			"count(CASE WHEN status = ? THEN 1 END) AS failed_removed",
+			database.DeliveryStatusFailed).
+		Where("event_id IN ?", ids).Group("target_id").Find(&removed),
+		"idx_deliveries_event_id (event_id=?)")
+	assertPlanUses(t, db, dry.Unscoped().Where("event_id IN ?", ids).
+		Delete(&database.Delivery{}), "idx_deliveries_event_id (event_id=?)")
+	assertPlanUses(t, db, dry.Unscoped().Where("id IN ?", ids).
+		Delete(&database.Event{}), "sqlite_autoindex_events_1 (id=?)")
+}
+
+// TestStatisticsQueriesUseTheirIndexes does the same for the webhook
+// page's statistics (readEventStats in the handlers): deliveries in
+// progress, each target's deliveries finished since a time, which must
+// come from the index alone, events received since a time, and the
+// newest event, which must come straight off an index rather than from
+// sorting every event.
+func TestStatisticsQueriesUseTheirIndexes(t *testing.T) {
+	t.Parallel()
+
+	mgr, lc := setupTestWebhookDBManager(t)
+	ctx := context.Background()
+	require.NoError(t, lc.Start(ctx))
+
+	defer func() { require.NoError(t, lc.Stop(ctx)) }()
+
+	db, err := mgr.GetDB(uuid.New().String())
+	require.NoError(t, err)
+
+	dry := db.Session(&gorm.Session{DryRun: true})
+	since := time.Now()
+
+	var (
+		count    int64
+		newest   []time.Time
+		byTarget []struct{ TargetID string }
+	)
+
+	assertPlanUses(t, db, dry.Model(&database.Delivery{}).
+		Where("status IN ?", []database.DeliveryStatus{
+			database.DeliveryStatusPending,
+			database.DeliveryStatusRetrying,
+		}).Count(&count),
+		"idx_deliveries_status (status=? AND deleted_at=?)")
+	assertPlanUses(t, db, dry.Model(&database.Delivery{}).
+		Select("target_id, "+
+			"count(CASE WHEN status = ? THEN 1 END) AS delivered, "+
+			"count(CASE WHEN status = ? THEN 1 END) AS failed",
+			database.DeliveryStatusDelivered,
+			database.DeliveryStatusFailed).
+		Where("status IN ? AND finished_at >= ?",
+			[]database.DeliveryStatus{
+				database.DeliveryStatusDelivered,
+				database.DeliveryStatusFailed,
+			}, since).
+		Group("target_id").Find(&byTarget),
+		"COVERING INDEX idx_deliveries_status "+
+			"(status=? AND deleted_at=? AND finished_at>?)")
+	assertPlanUses(t, db, dry.Model(&database.Event{}).
+		Where("created_at >= ?", since).Count(&count),
+		"idx_events_deleted_at_created_at "+
+			"(deleted_at=? AND created_at>?)")
+
+	newestEvent := dry.Model(&database.Event{}).
+		Order("created_at DESC").Limit(1).Pluck("created_at", &newest)
+	assertPlanUses(t, db, newestEvent,
+		"idx_events_deleted_at_created_at (deleted_at=?)")
+	assert.NotContains(t, queryPlan(t, db, newestEvent), "TEMP B-TREE")
 }
 
 // assertPlanUses asserts that SQLite's plan for a statement GORM built
@@ -152,6 +216,18 @@ func assertPlanUses(
 ) {
 	t.Helper()
 
+	plan := queryPlan(t, db, built)
+
+	for _, index := range indexes {
+		assert.Contains(t, plan, index, built.Statement.SQL.String())
+	}
+}
+
+// queryPlan returns SQLite's plan for a statement GORM built in a dry
+// run, run with the same SQL and arguments GORM would send.
+func queryPlan(t *testing.T, db, built *gorm.DB) string {
+	t.Helper()
+
 	var plan []struct{ Detail string }
 
 	require.NoError(t, db.Raw(
@@ -159,8 +235,5 @@ func assertPlanUses(
 		built.Statement.Vars...,
 	).Scan(&plan).Error)
 
-	for _, index := range indexes {
-		assert.Contains(t, fmt.Sprint(plan), index,
-			built.Statement.SQL.String())
-	}
+	return fmt.Sprint(plan)
 }

internal/database/export_test.go

@@ -28,6 +28,10 @@ func NewTestRetentionReaper(
 	}
 }
 
+// ExportReapBatchSize exposes how many expired events one retention
+// transaction deletes.
+const ExportReapBatchSize = reapBatchSize
+
 // ExportSweep runs a single retention sweep synchronously for tests.
 func (r *RetentionReaper) ExportSweep(ctx context.Context) {
 	r.sweep(ctx)

internal/database/model_delivery.go

@@ -1,6 +1,10 @@
 package database
 
-import "gorm.io/gorm"
+import (
+	"time"
+
+	"gorm.io/gorm"
+)
 
 // DeliveryStatus represents the status of a delivery
 type DeliveryStatus string
@@ -37,7 +41,7 @@ type Delivery struct {
 	BaseModel
 
 	EventID  string         `gorm:"type:uuid;not null;index:idx_deliveries_event_id,priority:1"       json:"eventId"`
-	TargetID string         `gorm:"type:uuid;not null"                                                json:"targetId"`
+	TargetID string         `gorm:"type:uuid;not null;index:idx_deliveries_status,priority:4"         json:"targetId"`
 	Status   DeliveryStatus `gorm:"not null;default:'pending';index:idx_deliveries_status,priority:1" json:"status"`
 
 	// DeletedAt repeats the BaseModel field only to be the second column
@@ -45,6 +49,13 @@ type Delivery struct {
 	// gives.
 	DeletedAt gorm.DeletedAt `gorm:"index:idx_deliveries_event_id,priority:2;index:idx_deliveries_status,priority:2" json:"deletedAt,omitzero"`
 
+	// FinishedAt is when the delivery became delivered or failed, and
+	// nil while it is pending or retrying. It and then TargetID end the
+	// status index, so the webhook page counts each target's deliveries
+	// that finished in a recent window by reading just that window from
+	// the index.
+	FinishedAt *time.Time `gorm:"index:idx_deliveries_status,priority:3" json:"finishedAt,omitempty"`
+
 	// Relations
 	Event           Event            `json:"event,omitzero"`
 	Target          Target           `json:"target,omitzero"`

internal/database/model_totals.go

@@ -0,0 +1,91 @@
+package database
+
+import (
+	"fmt"
+
+	"gorm.io/gorm"
+)
+
+// The running totals in a webhook's event database keep the webhook
+// page's lifetime figures right after retention has removed the rows
+// they count, and let the page show them without counting every row.
+// Each total changes in the transaction that writes or deletes the
+// rows it counts.
+
+// EventTotals is the single row counting a webhook's events: every
+// event ever stored, and how many of them retention has deleted.
+type EventTotals struct {
+	ID int64 `gorm:"primaryKey"`
+
+	Events        int64 `gorm:"not null"`
+	EventsRemoved int64 `gorm:"not null"`
+}
+
+// TableName names the table AddEventTotals updates.
+func (EventTotals) TableName() string {
+	return "event_totals"
+}
+
+// TargetTotals is one row per target counting its deliveries: every
+// delivery ever created, how many became delivered and how many
+// failed, and how many deliveries and failed deliveries retention has
+// deleted. The webhook's delivery figures are these rows summed.
+type TargetTotals struct {
+	TargetID string `gorm:"type:uuid;primaryKey"`
+
+	Deliveries int64 `gorm:"not null"`
+	Delivered  int64 `gorm:"not null"`
+	Failed     int64 `gorm:"not null"`
+
+	DeliveriesRemoved int64 `gorm:"not null"`
+	FailedRemoved     int64 `gorm:"not null"`
+}
+
+// TableName names the table AddTargetTotals updates.
+func (TargetTotals) TableName() string {
+	return "target_totals"
+}
+
+// AddEventTotals adds each count in add to the webhook's event totals.
+// Call it on the transaction that writes or deletes the events it
+// counts.
+func AddEventTotals(tx *gorm.DB, add EventTotals) error {
+	err := tx.Exec(
+		`UPDATE event_totals SET
+			events = events + ?,
+			events_removed = events_removed + ?`,
+		add.Events, add.EventsRemoved,
+	).Error
+	if err != nil {
+		return fmt.Errorf("adding to event totals: %w", err)
+	}
+
+	return nil
+}
+
+// AddTargetTotals adds each count in add to the totals of the target
+// add.TargetID names, creating its row the first time. Call it on the
+// transaction that writes or deletes the deliveries it counts.
+func AddTargetTotals(tx *gorm.DB, add TargetTotals) error {
+	err := tx.Exec(
+		`INSERT INTO target_totals (target_id, deliveries, delivered,
+			failed, deliveries_removed, failed_removed)
+		VALUES (?, ?, ?, ?, ?, ?)
+		ON CONFLICT (target_id) DO UPDATE SET
+			deliveries = deliveries + excluded.deliveries,
+			delivered = delivered + excluded.delivered,
+			failed = failed + excluded.failed,
+			deliveries_removed =
+				deliveries_removed + excluded.deliveries_removed,
+			failed_removed = failed_removed + excluded.failed_removed`,
+		add.TargetID, add.Deliveries, add.Delivered,
+		add.Failed, add.DeliveriesRemoved, add.FailedRemoved,
+	).Error
+	if err != nil {
+		return fmt.Errorf(
+			"adding to totals of target %s: %w", add.TargetID, err,
+		)
+	}
+
+	return nil
+}

internal/database/models.go

@@ -2,7 +2,8 @@ package database
 
 // Migrate runs database migrations for the main application database.
 // Only configuration-tier models are stored in the main database.
-// Event-tier models (Event, Delivery, DeliveryResult) live in
+// Event-tier models (Event, Delivery, DeliveryResult, EventTotals,
+// TargetTotals) live in
 // per-webhook dedicated databases managed by WebhookDBManager.
 func (d *Database) Migrate() error {
 	return d.db.AutoMigrate(

internal/database/retention.go

@@ -18,6 +18,19 @@ import (
 // computation.
 const hoursPerDay = 24
 
+// reapBatchSize is how many expired events one retention transaction
+// deletes. A transaction holds the event database's write lock, which
+// the receiver and the delivery workers wait for, so a large prune is
+// split into transactions each short enough to finish well inside the
+// busy timeout.
+const reapBatchSize = 1000
+
+// reapBatchPause is how long retention waits after one batch before
+// starting the next. A writer waiting for the write lock checks for it
+// again after at most 100 ms, so a longer pause lets it in between two
+// batches instead of only after the whole prune.
+const reapBatchPause = 200 * time.Millisecond
+
 // RetentionReaperParams holds the fx dependencies for the
 // RetentionReaper.
 type RetentionReaperParams struct {
@@ -265,57 +278,99 @@ func retentionCutoff(
 	), true
 }
 
-// reapExpired hard-deletes, in foreign-key-safe order, the delivery
-// results, deliveries, and events associated with events older than
-// cutoff. Deletes are unscoped so rows are physically removed rather
-// than soft-deleted, reclaiming disk. It returns the number of events
-// deleted.
+// reapExpired hard-deletes the events older than cutoff, with their
+// deliveries and delivery results, reapBatchSize events per
+// transaction with reapBatchPause between transactions, until none is
+// left. It returns the number of events deleted.
 func reapExpired(db *gorm.DB, cutoff time.Time) (int64, error) {
-	// Fresh subqueries are built per statement to avoid reusing a
-	// mutated builder across executions.
-	expiredEventIDs := func() *gorm.DB {
-		return db.Model(&Event{}).
-			Select("id").
-			Where("created_at < ?", cutoff)
-	}
-	expiredDeliveryIDs := func() *gorm.DB {
-		return db.Model(&Delivery{}).
-			Select("id").
-			Where("event_id IN (?)", expiredEventIDs())
-	}
+	var total int64
 
-	// 1. Delivery results whose delivery belongs to an expired event.
-	res := db.Unscoped().
-		Where("delivery_id IN (?)", expiredDeliveryIDs()).
-		Delete(&DeliveryResult{})
-	if res.Error != nil {
-		return 0, fmt.Errorf(
-			"deleting expired delivery results: %w",
-			res.Error,
-		)
-	}
+	for {
+		var eventIDs []string
 
-	// 2. Deliveries belonging to an expired event.
-	del := db.Unscoped().
-		Where("event_id IN (?)", expiredEventIDs()).
-		Delete(&Delivery{})
-	if del.Error != nil {
-		return 0, fmt.Errorf(
-			"deleting expired deliveries: %w",
-			del.Error,
-		)
-	}
-
-	// 3. The expired events themselves.
-	ev := db.Unscoped().
+		err := db.Transaction(func(tx *gorm.DB) error {
+			err := tx.Unscoped().Model(&Event{}).
 				Where("created_at < ?", cutoff).
-		Delete(&Event{})
-	if ev.Error != nil {
-		return 0, fmt.Errorf(
-			"deleting expired events: %w",
-			ev.Error,
-		)
+				Limit(reapBatchSize).
+				Pluck("id", &eventIDs).Error
+			if err != nil {
+				return fmt.Errorf("selecting expired events: %w", err)
 			}
 
-	return ev.RowsAffected, nil
+			if len(eventIDs) == 0 {
+				return nil
+			}
+
+			return deleteEvents(tx, eventIDs)
+		})
+		if err != nil {
+			return total, err
+		}
+
+		total += int64(len(eventIDs))
+
+		if len(eventIDs) < reapBatchSize {
+			return total, nil
+		}
+
+		time.Sleep(reapBatchPause)
+	}
+}
+
+// deleteEvents hard-deletes the given events and, in foreign-key-safe
+// order before them, their delivery results and deliveries, then adds
+// what it deleted to the running totals. It runs on reapExpired's
+// transaction, so the totals change exactly when the rows do. Deletes
+// are unscoped so rows are physically removed rather than
+// soft-deleted, reclaiming disk.
+func deleteEvents(tx *gorm.DB, eventIDs []string) error {
+	// 1. The delivery results of the events' deliveries.
+	err := tx.Unscoped().
+		Where("delivery_id IN (?)", tx.Unscoped().Model(&Delivery{}).
+			Select("id").
+			Where("event_id IN ?", eventIDs)).
+		Delete(&DeliveryResult{}).Error
+	if err != nil {
+		return fmt.Errorf("deleting expired delivery results: %w", err)
+	}
+
+	// 2. The events' deliveries, after counting them, and the failed
+	// ones among them, per target. The status is tested in the select
+	// list rather than the WHERE clause: there, SQLite would read every
+	// failed delivery the webhook has through the status index,
+	// instead of only these through the event_id index.
+	var removed []TargetTotals
+
+	err = tx.Unscoped().Model(&Delivery{}).
+		Select("target_id, count(*) AS deliveries_removed, "+
+			"count(CASE WHEN status = ? THEN 1 END) AS failed_removed",
+			DeliveryStatusFailed).
+		Where("event_id IN ?", eventIDs).
+		Group("target_id").
+		Find(&removed).Error
+	if err != nil {
+		return fmt.Errorf("counting expired deliveries: %w", err)
+	}
+
+	err = tx.Unscoped().
+		Where("event_id IN ?", eventIDs).
+		Delete(&Delivery{}).Error
+	if err != nil {
+		return fmt.Errorf("deleting expired deliveries: %w", err)
+	}
+
+	// 3. The events themselves.
+	ev := tx.Unscoped().Where("id IN ?", eventIDs).Delete(&Event{})
+	if ev.Error != nil {
+		return fmt.Errorf("deleting expired events: %w", ev.Error)
+	}
+
+	for i := range removed {
+		err = AddTargetTotals(tx, removed[i])
+		if err != nil {
+			return err
+		}
+	}
+
+	return AddEventTotals(tx, EventTotals{EventsRemoved: ev.RowsAffected})
 }

internal/database/totals_test.go

@@ -0,0 +1,305 @@
+package database_test
+
+import (
+	"context"
+	"net/http"
+	"testing"
+	"time"
+
+	"github.com/google/uuid"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"gorm.io/gorm"
+	"sneak.berlin/go/webhooker/internal/database"
+)
+
+// readEventTotals reads a webhook database's row of event totals,
+// asserting that it has exactly one.
+func readEventTotals(t *testing.T, db *gorm.DB) database.EventTotals {
+	t.Helper()
+
+	var rows []database.EventTotals
+
+	require.NoError(t, db.Find(&rows).Error)
+	require.Len(t, rows, 1)
+
+	return rows[0]
+}
+
+// readTargetTotals reads a webhook database's target totals, keyed by
+// target.
+func readTargetTotals(
+	t *testing.T, db *gorm.DB,
+) map[string]database.TargetTotals {
+	t.Helper()
+
+	var rows []database.TargetTotals
+
+	require.NoError(t, db.Find(&rows).Error)
+
+	byTarget := make(map[string]database.TargetTotals, len(rows))
+	for _, row := range rows {
+		byTarget[row.TargetID] = row
+	}
+
+	return byTarget
+}
+
+// TestWebhookDBManager_TotalsSurviveReopen verifies that a new event
+// database starts with one row of zero event totals and no target
+// totals, that adding to a target twice adds to the one row, and that
+// opening the database again keeps everything added.
+func TestWebhookDBManager_TotalsSurviveReopen(t *testing.T) {
+	t.Parallel()
+
+	mgr, lc := setupTestWebhookDBManager(t)
+	ctx := context.Background()
+	require.NoError(t, lc.Start(ctx))
+
+	defer func() { require.NoError(t, lc.Stop(ctx)) }()
+
+	webhookID := uuid.New().String()
+
+	db, err := mgr.GetDB(webhookID)
+	require.NoError(t, err)
+
+	fresh := readEventTotals(t, db)
+	assert.Equal(t, database.EventTotals{ID: fresh.ID}, fresh)
+	assert.Empty(t, readTargetTotals(t, db))
+
+	first, second := uuid.New().String(), uuid.New().String()
+
+	require.NoError(t, database.AddEventTotals(db, database.EventTotals{
+		Events: 2,
+	}))
+	require.NoError(t, database.AddTargetTotals(db, database.TargetTotals{
+		TargetID: first, Deliveries: 2, Delivered: 1,
+	}))
+	require.NoError(t, database.AddTargetTotals(db, database.TargetTotals{
+		TargetID: first, Failed: 1,
+	}))
+	require.NoError(t, database.AddTargetTotals(db, database.TargetTotals{
+		TargetID: second, Deliveries: 1,
+	}))
+
+	// Drop the cached connection so the next open reopens the file,
+	// as a restart would.
+	require.NoError(t, mgr.CloseAll())
+
+	db, err = mgr.GetDB(webhookID)
+	require.NoError(t, err)
+
+	assert.Equal(t, database.EventTotals{ID: fresh.ID, Events: 2},
+		readEventTotals(t, db))
+	assert.Equal(t, map[string]database.TargetTotals{
+		first: {
+			TargetID: first, Deliveries: 2, Delivered: 1, Failed: 1,
+		},
+		second: {TargetID: second, Deliveries: 1},
+	}, readTargetTotals(t, db))
+}
+
+// seedExpiredEvents stores count events created at the given time,
+// each with a delivered delivery to one target and a failed delivery
+// to the other, and one attempt for each delivery.
+func seedExpiredEvents(
+	t *testing.T,
+	db *gorm.DB,
+	webhookID string,
+	count int,
+	createdAt time.Time,
+	delivered, failed string,
+) {
+	t.Helper()
+
+	events := make([]database.Event, count)
+	deliveries := make([]database.Delivery, 0, 2*count)
+
+	for i := range events {
+		events[i] = database.Event{
+			WebhookID:    webhookID,
+			EntrypointID: uuid.New().String(),
+			Method:       http.MethodPost,
+		}
+		events[i].ID = uuid.New().String()
+		events[i].CreatedAt = createdAt
+
+		deliveries = append(deliveries,
+			database.Delivery{
+				EventID:  events[i].ID,
+				TargetID: delivered,
+				Status:   database.DeliveryStatusDelivered,
+			},
+			database.Delivery{
+				EventID:  events[i].ID,
+				TargetID: failed,
+				Status:   database.DeliveryStatusFailed,
+			},
+		)
+	}
+
+	require.NoError(t, db.CreateInBatches(events, 500).Error)
+	require.NoError(t, db.CreateInBatches(deliveries, 500).Error)
+
+	results := make([]database.DeliveryResult, len(deliveries))
+	for i := range deliveries {
+		results[i] = database.DeliveryResult{
+			DeliveryID: deliveries[i].ID, AttemptNum: 1,
+		}
+	}
+
+	require.NoError(t, db.CreateInBatches(results, 500).Error)
+}
+
+// TestRetentionReaper_PrunesMoreThanOneBatch verifies that a prune
+// larger than one transaction's batch removes every expired event with
+// its deliveries and delivery results, keeps the recent event, and
+// adds what it removed to the event and target totals, so the totals
+// within retention match the rows still stored.
+func TestRetentionReaper_PrunesMoreThanOneBatch(t *testing.T) {
+	t.Parallel()
+
+	env := setupRetentionTest(t)
+
+	webhookID := createWebhook(t, env.mainDB.DB(), 30)
+
+	db, err := env.mgr.GetDB(webhookID)
+	require.NoError(t, err)
+
+	expired := database.ExportReapBatchSize + 1
+	delivered, failed := uuid.New().String(), uuid.New().String()
+	seedExpiredEvents(t, db, webhookID, expired,
+		time.Now().Add(-40*24*time.Hour), delivered, failed)
+
+	// One recent event, delivered to the first target.
+	recent := seedEventChain(t, db, webhookID, time.Now())
+	require.NoError(t, db.Model(&database.Delivery{}).
+		Where("id = ?", recent.deliveryID).
+		Update("target_id", delivered).Error)
+
+	// The totals storing those rows would have left.
+	n := int64(expired)
+	require.NoError(t, database.AddEventTotals(db, database.EventTotals{
+		Events: n + 1,
+	}))
+	require.NoError(t, database.AddTargetTotals(db, database.TargetTotals{
+		TargetID: delivered, Deliveries: n + 1, Delivered: n + 1,
+	}))
+	require.NoError(t, database.AddTargetTotals(db, database.TargetTotals{
+		TargetID: failed, Deliveries: n, Failed: n,
+	}))
+
+	env.reaper.ExportSweep(context.Background())
+
+	// Only the recent event's rows are left.
+	for _, model := range []any{
+		&database.Event{}, &database.Delivery{}, &database.DeliveryResult{},
+	} {
+		var count int64
+
+		require.NoError(t, db.Model(model).Count(&count).Error)
+		assert.Equal(t, int64(1), count, "%T rows left", model)
+	}
+
+	assertChainPresent(t, db, recent)
+
+	eventTotals := readEventTotals(t, db)
+	assert.Equal(t, database.EventTotals{
+		ID: eventTotals.ID, Events: n + 1, EventsRemoved: n,
+	}, eventTotals)
+
+	targetTotals := readTargetTotals(t, db)
+	assert.Equal(t, map[string]database.TargetTotals{
+		delivered: {
+			TargetID: delivered, Deliveries: n + 1, Delivered: n + 1,
+			DeliveriesRemoved: n,
+		},
+		failed: {
+			TargetID: failed, Deliveries: n, Failed: n,
+			DeliveriesRemoved: n, FailedRemoved: n,
+		},
+	}, targetTotals)
+
+	// A sweep with nothing left to remove changes nothing.
+	env.reaper.ExportSweep(context.Background())
+
+	assert.Equal(t, eventTotals, readEventTotals(t, db))
+	assert.Equal(t, targetTotals, readTargetTotals(t, db))
+}
+
+// TestRetentionReaper_WriteDuringPruneSucceeds verifies that a prune
+// of several batches lets other writers in between its batches: an
+// event stored once the first batch is deleted is stored while expired
+// events are still left, not only after the prune has finished.
+func TestRetentionReaper_WriteDuringPruneSucceeds(t *testing.T) {
+	t.Parallel()
+
+	env := setupRetentionTest(t)
+
+	webhookID := createWebhook(t, env.mainDB.DB(), 30)
+
+	db, err := env.mgr.GetDB(webhookID)
+	require.NoError(t, err)
+
+	// Three batches of expired events, with nothing else stored: only
+	// the number of batches matters here.
+	expired := make([]database.Event, 3*database.ExportReapBatchSize)
+	for i := range expired {
+		expired[i] = database.Event{
+			WebhookID:    webhookID,
+			EntrypointID: uuid.New().String(),
+			Method:       http.MethodPost,
+		}
+		expired[i].CreatedAt = time.Now().Add(-40 * 24 * time.Hour)
+	}
+
+	require.NoError(t, db.CreateInBatches(expired, 500).Error)
+
+	cutoff := time.Now().Add(-30 * 24 * time.Hour)
+	countExpired := func() int64 {
+		var count int64
+
+		require.NoError(t, db.Model(&database.Event{}).
+			Where("created_at < ?", cutoff).
+			Count(&count).Error)
+
+		return count
+	}
+
+	pruned := make(chan struct{})
+
+	go func() {
+		defer close(pruned)
+
+		env.reaper.ExportSweep(context.Background())
+	}()
+
+	t.Cleanup(func() { <-pruned })
+
+	// Every stored event is expired until the write below.
+	require.Eventually(t, func() bool {
+		var count int64
+
+		err := db.Model(&database.Event{}).Count(&count).Error
+
+		return err == nil && count < int64(len(expired))
+	}, 10*time.Second, 10*time.Millisecond)
+
+	event := &database.Event{
+		WebhookID:    webhookID,
+		EntrypointID: uuid.New().String(),
+		Method:       http.MethodPost,
+	}
+	require.NoError(t, db.Create(event).Error)
+
+	assert.Positive(t, countExpired(),
+		"the event was stored only after the whole prune")
+
+	<-pruned
+
+	assert.Zero(t, countExpired())
+
+	var stored database.Event
+
+	require.NoError(t, db.First(&stored, "id = ?", event.ID).Error)
+}

internal/database/webhook_db_manager.go

@@ -35,7 +35,8 @@ var errInvalidCachedDBType = errors.New(
 
 // WebhookDBManager manages per-webhook SQLite database files
 // for event storage. Each webhook gets its own dedicated
-// database containing Events, Deliveries, and DeliveryResults.
+// database containing Events, Deliveries, DeliveryResults and the
+// running totals of them (EventTotals, TargetTotals).
 // Database connections are opened lazily and cached.
 type WebhookDBManager struct {
 	dataDir string
@@ -295,6 +296,7 @@ func (m *WebhookDBManager) openDB(
 	// Run migrations for event-tier models only
 	err = db.AutoMigrate(
 		&Event{}, &Delivery{}, &DeliveryResult{},
+		&EventTotals{}, &TargetTotals{},
 	)
 	if err != nil {
 		_ = sqlDB.Close()
@@ -305,6 +307,18 @@ func (m *WebhookDBManager) openDB(
 		)
 	}
 
+	// A new database gets its row of event totals, all zero. Target
+	// totals rows are created by the first delivery to each target.
+	err = db.FirstOrCreate(&EventTotals{}).Error
+	if err != nil {
+		_ = sqlDB.Close()
+
+		return nil, fmt.Errorf(
+			"creating event totals for webhook database %s: %w",
+			webhookID, err,
+		)
+	}
+
 	m.log.Info(
 		"opened per-webhook database",
 		"webhook_id", webhookID,

internal/delivery/delivery_totals_test.go

@@ -0,0 +1,116 @@
+package delivery_test
+
+import (
+	"testing"
+	"time"
+
+	"github.com/google/uuid"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"gorm.io/gorm"
+	"sneak.berlin/go/webhooker/internal/database"
+)
+
+// targetTotals reads one target's totals from a webhook database, all
+// zero when it has no row.
+func targetTotals(
+	t *testing.T, db *gorm.DB, targetID string,
+) database.TargetTotals {
+	t.Helper()
+
+	var rows []database.TargetTotals
+
+	require.NoError(t, db.Where("target_id = ?", targetID).
+		Find(&rows).Error)
+
+	if len(rows) == 0 {
+		return database.TargetTotals{TargetID: targetID}
+	}
+
+	return rows[0]
+}
+
+// TestUpdateDeliveryStatus_FinishTimeAndTargetTotals pins what a status
+// write records for the webhook page's statistics: the time a delivery
+// finished, set only when it becomes delivered or failed, and one more
+// on its target's delivered or failed total.
+func TestUpdateDeliveryStatus_FinishTimeAndTargetTotals(t *testing.T) {
+	t.Parallel()
+
+	tests := []struct {
+		status    database.DeliveryStatus
+		finished  bool
+		delivered int64
+		failed    int64
+	}{
+		{database.DeliveryStatusRetrying, false, 0, 0},
+		{database.DeliveryStatusDelivered, true, 1, 0},
+		{database.DeliveryStatusFailed, true, 0, 1},
+	}
+
+	for _, tt := range tests {
+		t.Run(string(tt.status), func(t *testing.T) {
+			t.Parallel()
+
+			db := testWebhookDB(t)
+			e := testEngine(t, 1)
+			event := seedEvent(t, db, `{}`)
+			targetID := uuid.New().String()
+			d := seedDelivery(
+				t, db, event.ID, targetID,
+				database.DeliveryStatusPending,
+			)
+
+			before := time.Now()
+
+			require.NoError(t, e.ExportUpdateDeliveryStatus(
+				db, &d, tt.status,
+			))
+
+			var stored database.Delivery
+
+			require.NoError(t, db.First(&stored, "id = ?", d.ID).Error)
+			assert.Equal(t, tt.status, stored.Status)
+
+			if tt.finished {
+				require.NotNil(t, stored.FinishedAt)
+				assert.False(t, stored.FinishedAt.Before(before))
+			} else {
+				assert.Nil(t, stored.FinishedAt)
+			}
+
+			assert.Equal(t, database.TargetTotals{
+				TargetID:  targetID,
+				Delivered: tt.delivered,
+				Failed:    tt.failed,
+			}, targetTotals(t, db, targetID))
+		})
+	}
+}
+
+// TestUpdateDeliveryStatus_DeletedDeliveryIsNotCounted covers a
+// delivery retention deleted while the engine still held it. Failing
+// it afterwards writes no row, so it adds no failure either: retention
+// has already counted what it removed.
+func TestUpdateDeliveryStatus_DeletedDeliveryIsNotCounted(t *testing.T) {
+	t.Parallel()
+
+	db := testWebhookDB(t)
+	e := testEngine(t, 1)
+	event := seedEvent(t, db, `{}`)
+	targetID := uuid.New().String()
+	d := seedDelivery(
+		t, db, event.ID, targetID,
+		database.DeliveryStatusRetrying,
+	)
+
+	require.NoError(t, db.Unscoped().
+		Delete(&database.Delivery{}, "id = ?", d.ID).Error)
+
+	require.NoError(t, e.ExportUpdateDeliveryStatus(
+		db, &d, database.DeliveryStatusFailed,
+	))
+
+	assert.Equal(t, database.TargetTotals{TargetID: targetID},
+		targetTotals(t, db, targetID))
+}

internal/delivery/engine.go

@@ -531,6 +531,11 @@ func (e *Engine) processRetryTask(
 		return
 	}
 
+	// Set before anything below can fail the delivery: the failure is
+	// added to this target's totals.
+	d.EventID = task.EventID
+	d.TargetID = task.TargetID
+
 	if d.Status != database.DeliveryStatusRetrying {
 		e.log.Debug(
 			"skipping retry for delivery "+
@@ -562,8 +567,6 @@ func (e *Engine) processRetryTask(
 	}
 
 	target := buildTargetFromTask(task)
-	d.EventID = task.EventID
-	d.TargetID = task.TargetID
 	d.Event = event
 	d.Target = target
 
@@ -1554,8 +1557,9 @@ func (e *Engine) updateDeliveryStatus(
 	targetType database.TargetType,
 	status database.DeliveryStatus,
 ) error {
-	err := webhookDB.Model(d).
-		Update("status", status).Error
+	err := webhookDB.Transaction(func(tx *gorm.DB) error {
+		return writeDeliveryStatus(tx, d, status)
+	})
 	if err != nil {
 		return fmt.Errorf(
 			"updating delivery %s to status %s: %w",
@@ -1574,6 +1578,36 @@ func (e *Engine) updateDeliveryStatus(
 	return nil
 }
 
+// writeDeliveryStatus writes a delivery's new status. A delivery that
+// becomes delivered or failed also gets the time it finished, and is
+// added to its target's delivered or failed total. It is counted only
+// if the row was still there to update: retention may have deleted it
+// while the engine was working on it.
+func writeDeliveryStatus(
+	tx *gorm.DB,
+	d *database.Delivery,
+	status database.DeliveryStatus,
+) error {
+	if !status.Terminal() {
+		return tx.Model(d).Update("status", status).Error
+	}
+
+	res := tx.Model(d).Updates(map[string]any{
+		"status":      status,
+		"finished_at": time.Now(),
+	})
+	if res.Error != nil || res.RowsAffected == 0 {
+		return res.Error
+	}
+
+	add := database.TargetTotals{TargetID: d.TargetID, Delivered: 1}
+	if status == database.DeliveryStatusFailed {
+		add = database.TargetTotals{TargetID: d.TargetID, Failed: 1}
+	}
+
+	return database.AddTargetTotals(tx, add)
+}
+
 // settleStatus moves a delivery to its outcome status and reports a
 // failed write through bookkeepingFailed, which leaves the row
 // recoverable. It exists so the target call sites read as one

internal/delivery/engine_test.go

@@ -57,7 +57,10 @@ func testWebhookDB(t *testing.T) *gorm.DB {
 		&database.Event{},
 		&database.Delivery{},
 		&database.DeliveryResult{},
+		&database.EventTotals{},
+		&database.TargetTotals{},
 	))
+	require.NoError(t, db.Create(&database.EventTotals{}).Error)
 
 	return db
 }

internal/delivery/export_test.go

@@ -150,6 +150,16 @@ func (e *Engine) ExportDeliverSlack(
 	)
 }
 
+// ExportUpdateDeliveryStatus exposes updateDeliveryStatus. It passes no
+// target type, so no metric moves.
+func (e *Engine) ExportUpdateDeliveryStatus(
+	webhookDB *gorm.DB,
+	d *database.Delivery,
+	status database.DeliveryStatus,
+) error {
+	return e.updateDeliveryStatus(webhookDB, d, "", status)
+}
+
 // ExportProcessNewTask exposes processNewTask.
 func (e *Engine) ExportProcessNewTask(
 	ctx context.Context, task *Task,

internal/delivery/terminal_state_test.go

@@ -418,6 +418,38 @@ func TestProcessRetryTask_TargetDeleted_MakesNoAttempt(
 	assert.Zero(t, s.Engine.ExportInflightHeld())
 }
 
+// TestProcessRetryTask_TargetDeleted_CountsFailureOnTarget verifies
+// that the failure of a retry abandoned because its target is gone is
+// added to that target's own totals, not to a row with no target.
+func TestProcessRetryTask_TargetDeleted_CountsFailureOnTarget(
+	t *testing.T,
+) {
+	t.Parallel()
+
+	s := newISetup(t)
+
+	var hits atomic.Int64
+
+	task, targetID := tRetryChainSetup(
+		t, s, "gone-counted", &hits,
+	)
+
+	require.NoError(t, s.MainDB.Delete(
+		&database.Target{}, "id = ?", targetID,
+	).Error)
+
+	s.Engine.ExportProcessRetryTask(
+		context.Background(), &task,
+	)
+
+	var rows []database.TargetTotals
+
+	require.NoError(t, s.WebhookDB.Find(&rows).Error)
+	assert.Equal(t, []database.TargetTotals{
+		{TargetID: targetID, Failed: 1},
+	}, rows)
+}
+
 // TestProcessRetryTask_TargetPresent_StillDelivers is the guard's
 // mutation check: a liveness check that refused every retry would pass
 // the test above and break every retry there is.

internal/handlers/auth.go

@@ -103,9 +103,10 @@ func (h *Handlers) renderLoginError(
 // The credential check runs BEFORE any rate-limit budget is
 // consulted, and only a failed check spends budget. That is what
 // keeps the single administrative path reachable: behind the reverse
-// proxy this deployment requires, with TRUSTED_PROXIES unset, every
-// client shares one bucket, so a limiter spent on arrival lets any
-// stranger deny the operator's own correct password indefinitely.
+// proxy this deployment requires, when TRUSTED_PROXIES does not cover
+// it, every client shares one bucket, so a limiter spent on arrival
+// lets any stranger deny the operator's own correct password
+// indefinitely.
 //
 // Verifying first means every login POST costs an Argon2id hash, so
 // the work is taken under a bounded number of verification slots.

internal/handlers/auth_test.go

@@ -25,7 +25,7 @@ const (
 
 	// sharedProxyPeer is the whole point of this file. Production is
 	// required to run behind a TLS-terminating reverse proxy, and
-	// TRUSTED_PROXIES defaults to empty, so every client — attacker
+	// when TRUSTED_PROXIES does not cover it every client — attacker
 	// and operator alike — reaches the process from the proxy's
 	// address and shares one rate-limit bucket. Both parties in
 	// these tests therefore use the same RemoteAddr.
@@ -115,11 +115,11 @@ func floodFailures(
 // done-criterion of https://git.eeqj.de/sneak/webhooker/issues/150.
 //
 // The attacker and the operator share one rate-limit bucket, because
-// behind the mandated reverse proxy with TRUSTED_PROXIES unset every
-// client keys on the proxy's address. The attacker floods the
-// operator's own username — a single-admin product has a predictable
-// one — far past the failure limit. The operator must still be able
-// to log in with the correct password.
+// behind the mandated reverse proxy, when TRUSTED_PROXIES does not
+// cover it, every client keys on the proxy's address. The attacker
+// floods the operator's own username — a single-admin product has a
+// predictable one — far past the failure limit. The operator must
+// still be able to log in with the correct password.
 //
 // This fails if credentials stop being verified ahead of the limiter.
 func TestLogin_StrangersFloodCannotLockOutTheOperator(t *testing.T) {

internal/handlers/delivery_replay.go

@@ -299,8 +299,9 @@ func countInFlightDeliveries(
 	return count, err
 }
 
-// createReplayDelivery writes the new pending delivery row and returns
-// the task that carries it to the delivery engine.
+// createReplayDelivery writes the new pending delivery row, adds it to
+// its target's totals in the same transaction, and returns the task
+// that carries it to the delivery engine.
 //
 // The row is written with associations omitted, and neither Event nor
 // Target is populated on it: GORM's SaveBeforeAssociations would
@@ -319,7 +320,16 @@ func createReplayDelivery(
 		Status:   database.DeliveryStatusPending,
 	}
 
-	err := webhookDB.Omit(clause.Associations).Create(dlv).Error
+	err := webhookDB.Transaction(func(tx *gorm.DB) error {
+		err := tx.Omit(clause.Associations).Create(dlv).Error
+		if err != nil {
+			return err
+		}
+
+		return database.AddTargetTotals(tx, database.TargetTotals{
+			TargetID: dlv.TargetID, Deliveries: 1,
+		})
+	})
 	if err != nil {
 		return delivery.Task{}, err
 	}

internal/handlers/export_test.go

@@ -4,7 +4,9 @@ import (
 	"html/template"
 	"log/slog"
 	"net/http"
+	"time"
 
+	"gorm.io/gorm"
 	"sneak.berlin/go/webhooker/internal/database"
 )
 
@@ -69,6 +71,29 @@ func (s *Handlers) LoadEventLogViewsForTest(
 	return views
 }
 
+// WebhookStatsForTest returns the figures the statistics pane on a
+// webhook's page shows, from the webhook's entrypoints and targets
+// loaded as that page loads them.
+func (s *Handlers) WebhookStatsForTest(webhookID string) *WebhookStats {
+	var entrypoints []database.Entrypoint
+
+	s.db.DB().Where("webhook_id = ?", webhookID).Find(&entrypoints)
+
+	var targets []database.Target
+
+	s.db.DB().Where("webhook_id = ?", webhookID).Find(&targets)
+
+	return s.loadWebhookStats(webhookID, entrypoints, targets)
+}
+
+// FinishedByTargetForTest exposes finishedByTarget for use in the
+// handlers_test package.
+func FinishedByTargetForTest(
+	webhookDB *gorm.DB, since time.Time,
+) ([]TargetFinished, error) {
+	return finishedByTarget(webhookDB, since)
+}
+
 // AddTemplateForTest registers a template under a page name so that
 // the handlers_test package can drive the render path with a
 // template of its own.

internal/handlers/handlers.go

@@ -91,18 +91,22 @@ type Handlers struct {
 
 // parsePageTemplate parses a page-specific template set from the
 // embedded FS. Each page template is combined with the shared
-// base, htmlheader, and navbar templates. The page file must be
-// listed first so that its root action ({{template "base" .}})
-// becomes the template set's entry point.
-func parsePageTemplate(pageFile string) *template.Template {
-	return template.Must(
-		template.ParseFS(
-			templates.Templates,
+// base, htmlheader, and navbar templates, and with any further files
+// the page includes. The page file must be listed first so that its
+// root action ({{template "base" .}}) becomes the template set's entry
+// point.
+func parsePageTemplate(
+	pageFile string, included ...string,
+) *template.Template {
+	files := append([]string{
 		pageFile,
 		"base.html",
 		"htmlheader.html",
 		"navbar.html",
-		),
+	}, included...)
+
+	return template.Must(
+		template.ParseFS(templates.Templates, files...),
 	)
 }
 
@@ -131,7 +135,7 @@ func New(
 		"profile.html":       parsePageTemplate("profile.html"),
 		"sources_list.html":  parsePageTemplate("sources_list.html"),
 		"sources_new.html":   parsePageTemplate("sources_new.html"),
-		"source_detail.html": parsePageTemplate("source_detail.html"),
+		"source_detail.html": parsePageTemplate("source_detail.html", "webhook_stats.html"),
 		"source_edit.html":   parsePageTemplate("source_edit.html"),
 		"source_logs.html":   parsePageTemplate("source_logs.html"),
 		"target_edit.html":   parsePageTemplate("target_edit.html"),

internal/handlers/source_management.go

@@ -457,6 +457,7 @@ func (h *Handlers) renderSourceDetail(
 		"Targets":     delivery.NewTargetViews(targets),
 		"Events":      events,
 		"BaseURL":     baseURL,
+		"Stats":       h.loadWebhookStats(webhook.ID, entrypoints, targets),
 	}
 
 	h.renderTemplate(w, r, "source_detail.html", data)

internal/handlers/webhook.go

@@ -253,11 +253,12 @@ func requestEventSource(
 	}
 }
 
-// createAndFanOut writes the event and one pending delivery per target
-// in a single transaction, then hands the tasks to the delivery
-// engine. It is the only path by which an event and its deliveries are
-// created, so a resubmitted event is retried, SSRF-guarded and
-// circuit-broken exactly as a received one is.
+// createAndFanOut writes the event and one pending delivery per target,
+// and adds them to the webhook's running totals, in a single
+// transaction, then hands the tasks to the delivery engine. It is the
+// only path by which an event and its deliveries are created, so a
+// resubmitted event is retried, SSRF-guarded and circuit-broken
+// exactly as a received one is.
 //
 // The tasks are returned as well as queued, so a caller can report how
 // many targets the event went to.
@@ -297,6 +298,13 @@ func (h *Handlers) createAndFanOut(
 		return nil, nil, err
 	}
 
+	err = database.AddEventTotals(tx, database.EventTotals{Events: 1})
+	if err != nil {
+		tx.Rollback()
+
+		return nil, nil, err
+	}
+
 	err = tx.Commit().Error
 	if err != nil {
 		return nil, nil, fmt.Errorf(
@@ -355,8 +363,9 @@ func (h *Handlers) finishWebhookResponse(
 }
 
 // buildDeliveryTasks creates one pending delivery per target in the
-// transaction and returns the tasks for the delivery engine. The
-// caller owns the transaction and rolls it back on error.
+// transaction, adds each to its target's totals, and returns the tasks
+// for the delivery engine. The caller owns the transaction and rolls
+// it back on error.
 func buildDeliveryTasks(
 	tx *gorm.DB,
 	event *database.Event,
@@ -380,6 +389,13 @@ func buildDeliveryTasks(
 			)
 		}
 
+		err = database.AddTargetTotals(tx, database.TargetTotals{
+			TargetID: targets[i].ID, Deliveries: 1,
+		})
+		if err != nil {
+			return nil, err
+		}
+
 		tasks = append(tasks, delivery.Task{
 			DeliveryID:   dlv.ID,
 			EventID:      event.ID,

internal/handlers/webhook_stats.go

@@ -0,0 +1,271 @@
+package handlers
+
+import (
+	"fmt"
+	"time"
+
+	"gorm.io/gorm"
+	"sneak.berlin/go/webhooker/internal/database"
+)
+
+// The spans of the two recent windows the statistics pane reports on:
+// the last 10 minutes and the last 24 hours.
+const (
+	shortWindow = 10 * time.Minute
+	longWindow  = 24 * time.Hour
+)
+
+// percent turns a fraction into a percentage.
+const percent = 100
+
+// WebhookStats holds the figures in the statistics pane at the top of
+// the webhook page.
+type WebhookStats struct {
+	Entrypoints       int
+	ActiveEntrypoints int
+	Targets           int
+	ActiveTargets     int
+
+	// Lifetime counts every event, delivery and failure the webhook
+	// has had, and WithinRetention those still stored.
+	Lifetime        Counts
+	WithinRetention Counts
+
+	// InProgress counts the deliveries still pending or retrying.
+	InProgress int64
+
+	// LastEventAt is when the newest stored event arrived, or nil when
+	// none is stored.
+	LastEventAt *time.Time
+
+	Last10Minutes RecentWindow
+	Last24Hours   RecentWindow
+}
+
+// Counts holds a number of events, of deliveries and of failed
+// deliveries.
+type Counts struct {
+	Events     int64
+	Deliveries int64
+	Failures   int64
+}
+
+// RecentWindow holds what happened in one recent window: the events
+// received in it, and the deliveries that became delivered or failed in
+// it.
+type RecentWindow struct {
+	Events    int64
+	Delivered int64
+	Failed    int64
+}
+
+// TargetFinished is how many of one target's deliveries became
+// delivered, and how many failed, in a recent window.
+type TargetFinished struct {
+	TargetID  string
+	Delivered int64
+	Failed    int64
+}
+
+// FailurePercent is the share of the deliveries finished in the window
+// that failed, or a dash when none finished. Deliveries still pending
+// or retrying are not counted either way.
+func (w RecentWindow) FailurePercent() string {
+	finished := w.Delivered + w.Failed
+	if finished == 0 {
+		return "—"
+	}
+
+	return fmt.Sprintf(
+		"%.1f%%", percent*float64(w.Failed)/float64(finished),
+	)
+}
+
+// loadWebhookStats gathers the figures for the statistics pane from the
+// webhook's entrypoints and targets, as the page has already loaded
+// them, and from its event database. It returns nil, and logs why, when
+// the event database cannot be read.
+func (h *Handlers) loadWebhookStats(
+	webhookID string,
+	entrypoints []database.Entrypoint,
+	targets []database.Target,
+) *WebhookStats {
+	stats := &WebhookStats{
+		Entrypoints: len(entrypoints),
+		Targets:     len(targets),
+	}
+
+	for i := range entrypoints {
+		if entrypoints[i].Active {
+			stats.ActiveEntrypoints++
+		}
+	}
+
+	for i := range targets {
+		if targets[i].Active {
+			stats.ActiveTargets++
+		}
+	}
+
+	// Opening an event database that does not exist would create it,
+	// and it would hold nothing to count.
+	if !h.dbMgr.DBExists(webhookID) {
+		return stats
+	}
+
+	webhookDB, err := h.dbMgr.GetDB(webhookID)
+	if err == nil {
+		err = readEventStats(webhookDB, time.Now(), stats)
+	}
+
+	if err != nil {
+		h.log.Error(
+			"failed to read webhook statistics",
+			"webhook_id", webhookID,
+			"error", err,
+		)
+
+		return nil
+	}
+
+	return stats
+}
+
+// readEventStats fills in the figures that come from the webhook's
+// event database. None of them reads every stored row: the totals are
+// one row for the events and one per target for the deliveries, and
+// every other figure is read from an index, over only the rows it
+// counts.
+func readEventStats(
+	db *gorm.DB, now time.Time, stats *WebhookStats,
+) error {
+	err := readTotals(db, stats)
+	if err != nil {
+		return err
+	}
+
+	err = db.Model(&database.Delivery{}).
+		Where("status IN ?", []database.DeliveryStatus{
+			database.DeliveryStatusPending,
+			database.DeliveryStatusRetrying,
+		}).
+		Count(&stats.InProgress).Error
+	if err != nil {
+		return fmt.Errorf("counting deliveries in progress: %w", err)
+	}
+
+	var newest []time.Time
+
+	err = db.Model(&database.Event{}).
+		Order("created_at DESC").
+		Limit(1).
+		Pluck("created_at", &newest).Error
+	if err != nil {
+		return fmt.Errorf("reading newest event time: %w", err)
+	}
+
+	if len(newest) > 0 {
+		stats.LastEventAt = &newest[0]
+	}
+
+	stats.Last10Minutes, err = readRecentWindow(
+		db, now.Add(-shortWindow),
+	)
+	if err != nil {
+		return err
+	}
+
+	stats.Last24Hours, err = readRecentWindow(
+		db, now.Add(-longWindow),
+	)
+
+	return err
+}
+
+// readTotals fills in the lifetime and within-retention figures from
+// the running totals: the events' row, and the targets' rows summed.
+func readTotals(db *gorm.DB, stats *WebhookStats) error {
+	var events database.EventTotals
+
+	err := db.Take(&events).Error
+	if err != nil {
+		return fmt.Errorf("reading event totals: %w", err)
+	}
+
+	var targets []database.TargetTotals
+
+	err = db.Find(&targets).Error
+	if err != nil {
+		return fmt.Errorf("reading target totals: %w", err)
+	}
+
+	stats.Lifetime.Events = events.Events
+	stats.WithinRetention.Events = events.Events - events.EventsRemoved
+
+	for _, t := range targets {
+		stats.Lifetime.Deliveries += t.Deliveries
+		stats.Lifetime.Failures += t.Failed
+		stats.WithinRetention.Deliveries += t.Deliveries - t.DeliveriesRemoved
+		stats.WithinRetention.Failures += t.Failed - t.FailedRemoved
+	}
+
+	return nil
+}
+
+// readRecentWindow counts the events received, and the deliveries that
+// became delivered or failed, since the given time.
+func readRecentWindow(
+	db *gorm.DB, since time.Time,
+) (RecentWindow, error) {
+	var w RecentWindow
+
+	err := db.Model(&database.Event{}).
+		Where("created_at >= ?", since).
+		Count(&w.Events).Error
+	if err != nil {
+		return w, fmt.Errorf("counting recent events: %w", err)
+	}
+
+	byTarget, err := finishedByTarget(db, since)
+	if err != nil {
+		return w, err
+	}
+
+	for _, f := range byTarget {
+		w.Delivered += f.Delivered
+		w.Failed += f.Failed
+	}
+
+	return w, nil
+}
+
+// finishedByTarget counts, for each target, the deliveries that became
+// delivered and those that failed since the given time, in one query
+// over just that window of the deliveries' status index. A target with
+// neither is left out.
+func finishedByTarget(
+	db *gorm.DB, since time.Time,
+) ([]TargetFinished, error) {
+	var byTarget []TargetFinished
+
+	err := db.Model(&database.Delivery{}).
+		Select("target_id, "+
+			"count(CASE WHEN status = ? THEN 1 END) AS delivered, "+
+			"count(CASE WHEN status = ? THEN 1 END) AS failed",
+			database.DeliveryStatusDelivered,
+			database.DeliveryStatusFailed).
+		Where("status IN ? AND finished_at >= ?",
+			[]database.DeliveryStatus{
+				database.DeliveryStatusDelivered,
+				database.DeliveryStatusFailed,
+			}, since).
+		Group("target_id").
+		Find(&byTarget).Error
+	if err != nil {
+		return nil, fmt.Errorf(
+			"counting deliveries finished by target: %w", err,
+		)
+	}
+
+	return byTarget, nil
+}

internal/handlers/webhook_stats_test.go

@@ -0,0 +1,474 @@
+package handlers_test
+
+import (
+	"net/http"
+	"regexp"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/google/uuid"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"go.uber.org/fx/fxtest"
+	"gorm.io/gorm"
+	"gorm.io/gorm/clause"
+	"sneak.berlin/go/webhooker/internal/config"
+	"sneak.berlin/go/webhooker/internal/database"
+	"sneak.berlin/go/webhooker/internal/handlers"
+	"sneak.berlin/go/webhooker/internal/logger"
+	"sneak.berlin/go/webhooker/internal/session"
+)
+
+// statsEntrypoint adds an entrypoint to a webhook and returns its path.
+func statsEntrypoint(
+	t *testing.T, db *database.Database, webhookID string, active bool,
+) string {
+	t.Helper()
+
+	ep := &database.Entrypoint{
+		WebhookID: webhookID,
+		Path:      uuid.New().String(),
+	}
+
+	require.NoError(t, db.DB().Omit(clause.Associations).Create(ep).Error)
+	require.NoError(t, db.DB().Model(ep).Update("active", active).Error)
+
+	return ep.Path
+}
+
+// statsDelivery returns an event's delivery to a target.
+func statsDelivery(
+	t *testing.T, webhookDB *gorm.DB, eventID, targetID string,
+) database.Delivery {
+	t.Helper()
+
+	var d database.Delivery
+
+	require.NoError(t, webhookDB.Where(
+		"event_id = ? AND target_id = ?", eventID, targetID,
+	).First(&d).Error)
+
+	return d
+}
+
+// statsFinish settles a delivery as the delivery engine does: its
+// final status and the time it finished, and one more on its target's
+// delivered or failed total, in one transaction.
+func statsFinish(
+	t *testing.T,
+	webhookDB *gorm.DB,
+	d database.Delivery,
+	status database.DeliveryStatus,
+	at time.Time,
+) {
+	t.Helper()
+
+	add := database.TargetTotals{TargetID: d.TargetID, Delivered: 1}
+	if status == database.DeliveryStatusFailed {
+		add = database.TargetTotals{TargetID: d.TargetID, Failed: 1}
+	}
+
+	require.NoError(t, webhookDB.Transaction(func(tx *gorm.DB) error {
+		err := tx.Model(&database.Delivery{}).
+			Where("id = ?", d.ID).
+			Updates(map[string]any{"status": status, "finished_at": at}).
+			Error
+		if err != nil {
+			return err
+		}
+
+		return database.AddTargetTotals(tx, add)
+	}))
+}
+
+// statsAge moves an event's arrival back to the given time.
+func statsAge(
+	t *testing.T, webhookDB *gorm.DB, eventID string, at time.Time,
+) {
+	t.Helper()
+
+	require.NoError(t, webhookDB.Model(&database.Event{}).
+		Where("id = ?", eventID).
+		Update("created_at", at).Error)
+}
+
+// statsTargetTotals reads a webhook database's target totals, keyed by
+// target.
+func statsTargetTotals(
+	t *testing.T, webhookDB *gorm.DB,
+) map[string]database.TargetTotals {
+	t.Helper()
+
+	var rows []database.TargetTotals
+
+	require.NoError(t, webhookDB.Find(&rows).Error)
+
+	byTarget := make(map[string]database.TargetTotals, len(rows))
+	for _, row := range rows {
+		byTarget[row.TargetID] = row
+	}
+
+	return byTarget
+}
+
+// statsHistory is the webhook seedStatsHistory builds: its event
+// database, its newest event, and its two active targets.
+type statsHistory struct {
+	webhook       *database.Webhook
+	webhookDB     *gorm.DB
+	newest        database.Event
+	first, second string
+}
+
+// seedStatsHistory builds the webhook the statistics test checks: one
+// day of retention, two entrypoints (one inactive) and three targets
+// (one inactive). Three events arrive through the receiver, and so
+// each has a delivery to the two active targets. The oldest event is
+// past retention, the middle one six hours old, the newest just in.
+// Their deliveries are settled as the delivery engine would, and a
+// replay adds a pending delivery to the oldest event.
+func seedStatsHistory(
+	t *testing.T,
+	h *handlers.Handlers,
+	sess *session.Session,
+	db *database.Database,
+	dbMgr *database.WebhookDBManager,
+) statsHistory {
+	t.Helper()
+
+	wh := &database.Webhook{
+		UserID: deleteTestUserID, Name: "stats", RetentionDays: 1,
+	}
+	require.NoError(t, db.DB().Omit(clause.Associations).Create(wh).Error)
+
+	path := statsEntrypoint(t, db, wh.ID, true)
+	statsEntrypoint(t, db, wh.ID, false)
+
+	first := seedConfiguredTarget(
+		t, db, wh.ID, database.TargetTypeHTTP,
+		`{"url":"`+replayTargetURL+`"}`,
+	)
+	second := seedTarget(t, db, wh.ID, database.TargetTypeLog)
+	inactive := seedTarget(t, db, wh.ID, database.TargetTypeLog)
+	require.NoError(t, db.DB().Model(inactive).
+		Update("active", false).Error)
+
+	router := receiverRouter(h)
+
+	for range 3 {
+		require.Equal(t, http.StatusOK, postReceiver(t, router, path))
+	}
+
+	webhookDB, err := dbMgr.GetDB(wh.ID)
+	require.NoError(t, err)
+
+	events := listEvents(t, webhookDB)
+	require.Len(t, events, 3)
+
+	oldest, middle, newest := events[0], events[1], events[2]
+	now := time.Now()
+
+	statsAge(t, webhookDB, oldest.ID, now.Add(-50*time.Hour))
+	statsAge(t, webhookDB, middle.ID, now.Add(-6*time.Hour))
+
+	oldestFailure := statsDelivery(t, webhookDB, oldest.ID, first.ID)
+	statsFinish(t, webhookDB, oldestFailure,
+		database.DeliveryStatusFailed, now.Add(-49*time.Hour))
+	statsFinish(t, webhookDB,
+		statsDelivery(t, webhookDB, oldest.ID, second.ID),
+		database.DeliveryStatusDelivered, now.Add(-49*time.Hour))
+	statsFinish(t, webhookDB,
+		statsDelivery(t, webhookDB, middle.ID, first.ID),
+		database.DeliveryStatusFailed, now.Add(-5*time.Hour))
+	statsFinish(t, webhookDB,
+		statsDelivery(t, webhookDB, middle.ID, second.ID),
+		database.DeliveryStatusFailed, now.Add(-time.Minute))
+	statsFinish(t, webhookDB,
+		statsDelivery(t, webhookDB, newest.ID, first.ID),
+		database.DeliveryStatusDelivered, now.Add(-2*time.Minute))
+
+	require.Equal(t, http.StatusSeeOther,
+		postReplay(t, h, sess, wh.ID, oldestFailure.ID).Code)
+
+	return statsHistory{
+		webhook:   wh,
+		webhookDB: webhookDB,
+		newest:    newest,
+		first:     first.ID,
+		second:    second.ID,
+	}
+}
+
+// statsPrune runs the real retention reaper until it has removed one
+// event from the webhook's database, then stops it.
+func statsPrune(
+	t *testing.T,
+	db *database.Database,
+	dbMgr *database.WebhookDBManager,
+	log *logger.Logger,
+	webhookDB *gorm.DB,
+) {
+	t.Helper()
+
+	lc := fxtest.NewLifecycle(t)
+	database.NewRetentionReaper(lc, database.RetentionReaperParams{
+		Config: &config.Config{
+			RetentionSweepInterval: 10 * time.Millisecond,
+		},
+		Database:  db,
+		DBManager: dbMgr,
+		Logger:    log,
+	})
+
+	lc.RequireStart()
+
+	require.Eventually(t, func() bool {
+		var totals database.EventTotals
+
+		err := webhookDB.Take(&totals).Error
+
+		return err == nil && totals.EventsRemoved == 1
+	}, 10*time.Second, 10*time.Millisecond)
+
+	lc.RequireStop()
+}
+
+// statsPane returns the text of the statistics pane in a rendered
+// webhook page, everything from its heading to the next heading on the
+// page, with the markup taken out and each run of space made one
+// space. A table then reads header by header and row by row, each
+// row's label followed by its figures in column order.
+func statsPane(t *testing.T, page string) string {
+	t.Helper()
+
+	_, pane, found := strings.Cut(page, ">Statistics</h2>")
+	require.True(t, found, "the page has no statistics pane")
+
+	pane, _, _ = strings.Cut(pane, "<h2")
+	pane = regexp.MustCompile(`<[^>]*>`).ReplaceAllString(pane, " ")
+
+	return strings.Join(strings.Fields(pane), " ")
+}
+
+// assertStatsTargets checks, for the history seedStatsHistory builds,
+// each target's totals and its deliveries finished in the last 24
+// hours. The first target has three deliveries and the replay, the
+// second three; the inactive target has none and so no row.
+func assertStatsTargets(t *testing.T, hist statsHistory) {
+	t.Helper()
+
+	first, second := hist.first, hist.second
+
+	assert.Equal(t, map[string]database.TargetTotals{
+		first: {TargetID: first, Deliveries: 4, Delivered: 1, Failed: 2},
+		second: {
+			TargetID: second, Deliveries: 3, Delivered: 1, Failed: 1,
+		},
+	}, statsTargetTotals(t, hist.webhookDB))
+
+	lastDay, err := handlers.FinishedByTargetForTest(
+		hist.webhookDB, time.Now().Add(-24*time.Hour),
+	)
+	require.NoError(t, err)
+	assert.ElementsMatch(t, []handlers.TargetFinished{
+		{TargetID: first, Delivered: 1, Failed: 1},
+		{TargetID: second, Failed: 1},
+	}, lastDay)
+}
+
+// assertStatsPaneAfterPrune checks the rendered statistics pane for the
+// history seedStatsHistory builds, once retention has removed the
+// oldest event: each figure after its label, in its column.
+func assertStatsPaneAfterPrune(
+	t *testing.T,
+	h *handlers.Handlers,
+	sess *session.Session,
+	hist statsHistory,
+) {
+	t.Helper()
+
+	pane := statsPane(t, renderSourceDetailPage(t, h, sess, hist.webhook.ID))
+	lastEvent := hist.newest.CreatedAt.Format("2006-01-02 15:04:05 UTC")
+
+	assert.Contains(t, pane, "Entrypoints 2 (1 active) "+
+		"Targets 3 (2 active) "+
+		"Deliveries in progress 1 "+
+		"Last event "+lastEvent+" "+
+		"Retention 1 day")
+	assert.Contains(t, pane, "Lifetime Within retention "+
+		"Events 3 2 "+
+		"Deliveries 7 4 "+
+		"Failures 3 2")
+	assert.Contains(t, pane, "Last 10 minutes Last 24 hours "+
+		"Events 1 2 "+
+		"Failures 1 2 "+
+		"Failure percentage 50.0% 66.7%")
+}
+
+// TestWebhookStats_EveryFigureAcrossRetentionPrune checks every figure
+// the statistics pane shows for the history seedStatsHistory builds,
+// and each target's totals and recent figures, before and after the
+// real retention reaper removes the oldest event.
+func TestWebhookStats_EveryFigureAcrossRetentionPrune(t *testing.T) {
+	t.Parallel()
+
+	var (
+		h     *handlers.Handlers
+		sess  *session.Session
+		db    *database.Database
+		dbMgr *database.WebhookDBManager
+		log   *logger.Logger
+	)
+
+	app := newTestApp(t, &h, &sess, &db, &dbMgr, &log)
+	app.RequireStart()
+
+	t.Cleanup(app.RequireStop)
+
+	hist := seedStatsHistory(t, h, sess, db, dbMgr)
+	first, second := hist.first, hist.second
+
+	stats := h.WebhookStatsForTest(hist.webhook.ID)
+	require.NotNil(t, stats)
+
+	assert.Equal(t, 2, stats.Entrypoints)
+	assert.Equal(t, 1, stats.ActiveEntrypoints)
+	assert.Equal(t, 3, stats.Targets)
+	assert.Equal(t, 2, stats.ActiveTargets)
+	assert.Equal(t, handlers.Counts{Events: 3, Deliveries: 7, Failures: 3},
+		stats.Lifetime)
+	assert.Equal(t, stats.Lifetime, stats.WithinRetention)
+	assert.Equal(t, int64(2), stats.InProgress)
+	require.NotNil(t, stats.LastEventAt)
+	assert.True(t, hist.newest.CreatedAt.Equal(*stats.LastEventAt))
+	assert.Equal(t, handlers.RecentWindow{
+		Events: 1, Delivered: 1, Failed: 1,
+	}, stats.Last10Minutes)
+	assert.Equal(t, handlers.RecentWindow{
+		Events: 2, Delivered: 1, Failed: 2,
+	}, stats.Last24Hours)
+	assert.Equal(t, "50.0%", stats.Last10Minutes.FailurePercent())
+	assert.Equal(t, "66.7%", stats.Last24Hours.FailurePercent())
+
+	assertStatsTargets(t, hist)
+
+	// Retention removes the oldest event with its three deliveries:
+	// the first target's failed one and the pending replay, and the
+	// second target's delivered one.
+	statsPrune(t, db, dbMgr, log, hist.webhookDB)
+
+	after := h.WebhookStatsForTest(hist.webhook.ID)
+	require.NotNil(t, after)
+
+	assert.Equal(t, stats.Lifetime, after.Lifetime)
+	assert.Equal(t, handlers.Counts{Events: 2, Deliveries: 4, Failures: 2},
+		after.WithinRetention)
+	assert.Equal(t, int64(1), after.InProgress)
+	assert.Equal(t, stats.LastEventAt, after.LastEventAt)
+	assert.Equal(t, stats.Last10Minutes, after.Last10Minutes)
+	assert.Equal(t, stats.Last24Hours, after.Last24Hours)
+
+	assert.Equal(t, map[string]database.TargetTotals{
+		first: {
+			TargetID: first, Deliveries: 4, Delivered: 1, Failed: 2,
+			DeliveriesRemoved: 2, FailedRemoved: 1,
+		},
+		second: {
+			TargetID: second, Deliveries: 3, Delivered: 1, Failed: 1,
+			DeliveriesRemoved: 1,
+		},
+	}, statsTargetTotals(t, hist.webhookDB))
+
+	assertStatsPaneAfterPrune(t, h, sess, hist)
+}
+
+// TestWebhookStats_PaneShowsRetentionPeriod checks that the statistics
+// pane itself, not only the line at the foot of the page, shows the
+// webhook's retention period, for a finite one and for forever.
+func TestWebhookStats_PaneShowsRetentionPeriod(t *testing.T) {
+	t.Parallel()
+
+	var (
+		h    *handlers.Handlers
+		sess *session.Session
+		db   *database.Database
+	)
+
+	app := newTestApp(t, &h, &sess, &db)
+	app.RequireStart()
+
+	t.Cleanup(app.RequireStop)
+
+	tests := []struct {
+		retentionDays int
+		want          string
+	}{
+		{30, "30 days"},
+		{database.RetentionForeverDays, "forever"},
+	}
+
+	for _, tt := range tests {
+		wh := &database.Webhook{
+			UserID:        deleteTestUserID,
+			Name:          "retention",
+			RetentionDays: tt.retentionDays,
+		}
+		require.NoError(t,
+			db.DB().Omit(clause.Associations).Create(wh).Error)
+
+		pane := statsPane(t, renderSourceDetailPage(t, h, sess, wh.ID))
+		assert.Contains(t, pane, "Retention "+tt.want)
+	}
+}
+
+// TestWebhookStats_WebhookWithNoEvents covers a webhook whose event
+// database has never been opened: every count is zero, the
+// percentages are a dash, and showing the page does not create the
+// database.
+func TestWebhookStats_WebhookWithNoEvents(t *testing.T) {
+	t.Parallel()
+
+	var (
+		h     *handlers.Handlers
+		sess  *session.Session
+		db    *database.Database
+		dbMgr *database.WebhookDBManager
+	)
+
+	app := newTestApp(t, &h, &sess, &db, &dbMgr)
+	app.RequireStart()
+
+	t.Cleanup(app.RequireStop)
+
+	wh := seedWebhook(t, db)
+
+	assert.Equal(t, &handlers.WebhookStats{}, h.WebhookStatsForTest(wh.ID))
+	assert.Equal(t, "—", handlers.RecentWindow{}.FailurePercent())
+
+	pane := statsPane(t, renderSourceDetailPage(t, h, sess, wh.ID))
+	assert.Contains(t, pane, "Last event none")
+	assert.Contains(t, pane, "Failure percentage — —")
+	assert.False(t, dbMgr.DBExists(wh.ID))
+}
+
+// TestRecentWindow_FailurePercent pins the percentage: failed
+// deliveries out of all that finished in the window.
+func TestRecentWindow_FailurePercent(t *testing.T) {
+	t.Parallel()
+
+	tests := []struct {
+		window handlers.RecentWindow
+		want   string
+	}{
+		{handlers.RecentWindow{}, "—"},
+		{handlers.RecentWindow{Events: 4}, "—"},
+		{handlers.RecentWindow{Delivered: 3, Failed: 1}, "25.0%"},
+		{handlers.RecentWindow{Failed: 2}, "100.0%"},
+		{handlers.RecentWindow{Delivered: 2}, "0.0%"},
+	}
+
+	for _, tt := range tests {
+		assert.Equal(t, tt.want, tt.window.FailurePercent(), tt.window)
+	}
+}

internal/middleware/loginguard.go

@@ -108,10 +108,10 @@ type failureWindow struct {
 //
 // A limiter that spends budget on arrival cannot protect a
 // single-admin product: behind the reverse proxy the deployment
-// requires, with TRUSTED_PROXIES unset, every client keys on the
-// proxy, so a stranger trickling five POSTs a minute keeps the one
-// bucket full and the operator's own correct password is answered 429
-// forever. There is no second administrative path.
+// requires, when TRUSTED_PROXIES does not cover it, every client
+// keys on the proxy, so a stranger trickling five POSTs a minute
+// keeps the one bucket full and the operator's own correct password
+// is answered 429 forever. There is no second administrative path.
 //
 // So budget is spent only by a FAILED verification. A correct
 // password is never throttled, whatever the counters say, which is

internal/middleware/ratelimit.go

@@ -123,9 +123,8 @@ func bucketKey(addr netip.Addr) string {
 	return prefix.String()
 }
 
-// isTrustedProxy reports whether addr belongs to a network the
-// operator listed in TRUSTED_PROXIES. The list is empty by default,
-// so by default nothing is trusted.
+// isTrustedProxy reports whether addr belongs to a network in
+// TRUSTED_PROXIES, which by default is the RFC 1918 private ranges.
 func (m *Middleware) isTrustedProxy(addr netip.Addr) bool {
 	for _, prefix := range m.params.Config.TrustedProxies {
 		if prefix.Contains(addr) {

internal/middleware/ratelimit_test.go

@@ -384,8 +384,8 @@ const (
 	// trustedProxyCIDR is the proxy network the forwarded-path
 	// tests configure, and trustedPeer an address inside it. A
 	// production deployment is required to run behind a reverse
-	// proxy with TRUSTED_PROXIES set, so this is the shape the
-	// bucketing has to hold in.
+	// proxy that TRUSTED_PROXIES covers, either by the default or by
+	// a set value, so this is the shape the bucketing has to hold in.
 	trustedProxyCIDR = "10.0.0.0/8"
 	trustedPeer      = "10.0.0.1:44444"
 )
@@ -426,8 +426,8 @@ func assertSharedBucket(
 }
 
 // TestRateLimitKey_SpoofedForwardedFromUntrustedPeer is the test
-// this gating exists for: with no trusted proxies configured (the
-// default), a client that rotates a forwarded header on every
+// this gating exists for: from a peer that is not a trusted
+// proxy, a client that rotates a forwarded header on every
 // request must stay in one bucket. If forwarded headers were
 // trusted unconditionally, each spoofed value would mint a fresh
 // bucket and the limit would stop no one.
@@ -1097,8 +1097,9 @@ func TestPostRateLimit_IPv4IndependentPerAddress(t *testing.T) {
 // that arrives from trustedPeer — a configured trusted proxy — and
 // names forwarded as its client in X-Forwarded-For. That is the
 // production path: a deployment is required to run behind a reverse
-// proxy with TRUSTED_PROXIES set, so the forwarded address, not the
-// peer, is what the limiters bucket on there.
+// proxy that TRUSTED_PROXIES covers, either by the default or by a
+// set value, so the forwarded address, not the peer, is what the
+// limiters bucket on there.
 func forwardedKeyFor(
 	t *testing.T, m *middleware.Middleware, forwarded string,
 ) string {
@@ -1178,9 +1179,9 @@ func TestRateLimitKey_ForwardedIPv6BucketsByPrefix(t *testing.T) {
 //
 // Every existing test of this fallback uses an IPv4 proxy, where
 // bucketKey is the identity function, so replacing the call with
-// peer.String() leaves the whole suite green. Only operator-listed
-// addresses reach this line and the fallback is fail-closed, so this
-// pins behaviour rather than fixing a defect.
+// peer.String() leaves the whole suite green. Only addresses inside
+// TRUSTED_PROXIES reach this line and the fallback is fail-closed, so
+// this pins behaviour rather than fixing a defect.
 func TestRateLimitKey_TrustedPeerUnusableForwardedMasksPeer(
 	t *testing.T,
 ) {

internal/server/routes.go

@@ -154,11 +154,12 @@ func (s *Server) setupPageRoutes() {
 		r.Use(s.mw.NoCache())
 
 		// The login POST carries no pre-emptive rate limiter. Behind
-		// the reverse proxy production requires, with TRUSTED_PROXIES
-		// unset, every client shares one bucket, so a limiter spent
-		// on arrival lets any stranger deny the operator the only
-		// administrative path. The handler verifies credentials first
-		// and charges only failures; see Handlers.authenticateUser.
+		// the reverse proxy production requires, when TRUSTED_PROXIES
+		// does not cover it, every client shares one bucket, so a
+		// limiter spent on arrival lets any stranger deny the operator
+		// the only administrative path. The handler verifies
+		// credentials first and charges only failures; see
+		// Handlers.authenticateUser.
 		r.Get("/login", s.h.HandleLoginPage())
 		r.Post("/login", s.h.HandleLoginSubmit())
 

templates/source_detail.html

@@ -24,6 +24,8 @@
         </div>
     </div>
 
+    {{template "webhook_stats" .}}
+
     <div class="grid grid-cols-1 lg:grid-cols-2 gap-6">
         <!-- Entrypoints -->
         <div class="card">

templates/webhook_stats.html

@@ -0,0 +1,93 @@
+{{define "webhook_stats"}}
+<!-- Statistics pane at the top of the webhook page. -->
+<div class="card mb-6">
+    <div class="p-4 border-b border-gray-200">
+        <h2 class="text-lg font-medium text-gray-900">Statistics</h2>
+    </div>
+    {{with .Stats}}
+    <div class="p-4 flex flex-wrap gap-6 text-sm border-b border-gray-200">
+        <div>
+            <span class="text-gray-500">Entrypoints</span>
+            <span class="font-medium text-gray-900">{{.Entrypoints}}</span>
+            <span class="text-gray-500">({{.ActiveEntrypoints}} active)</span>
+        </div>
+        <div>
+            <span class="text-gray-500">Targets</span>
+            <span class="font-medium text-gray-900">{{.Targets}}</span>
+            <span class="text-gray-500">({{.ActiveTargets}} active)</span>
+        </div>
+        <div>
+            <span class="text-gray-500">Deliveries in progress</span>
+            <span class="font-medium text-gray-900">{{.InProgress}}</span>
+        </div>
+        <div>
+            <span class="text-gray-500">Last event</span>
+            <span class="font-medium text-gray-900">{{with .LastEventAt}}{{.Format "2006-01-02 15:04:05 UTC"}}{{else}}none{{end}}</span>
+        </div>
+        <div>
+            <span class="text-gray-500">Retention</span>
+            <span class="font-medium text-gray-900">{{$.Webhook.RetentionLabel}}</span>
+        </div>
+    </div>
+    <div class="p-4 grid grid-cols-1 lg:grid-cols-2 gap-6 text-sm">
+        <table class="w-full text-center text-gray-900">
+            <thead>
+                <tr class="border-b border-gray-200 text-xs text-gray-500 uppercase tracking-wide">
+                    <th></th>
+                    <th class="py-2 font-medium">Lifetime</th>
+                    <th class="py-2 font-medium">Within retention</th>
+                </tr>
+            </thead>
+            <tbody>
+                <tr>
+                    <td class="py-2 text-left text-gray-600">Events</td>
+                    <td class="py-2">{{.Lifetime.Events}}</td>
+                    <td class="py-2">{{.WithinRetention.Events}}</td>
+                </tr>
+                <tr>
+                    <td class="py-2 text-left text-gray-600">Deliveries</td>
+                    <td class="py-2">{{.Lifetime.Deliveries}}</td>
+                    <td class="py-2">{{.WithinRetention.Deliveries}}</td>
+                </tr>
+                <tr>
+                    <td class="py-2 text-left text-gray-600">Failures</td>
+                    <td class="py-2">{{.Lifetime.Failures}}</td>
+                    <td class="py-2">{{.WithinRetention.Failures}}</td>
+                </tr>
+            </tbody>
+        </table>
+        <div>
+            <table class="w-full text-center text-gray-900">
+                <thead>
+                    <tr class="border-b border-gray-200 text-xs text-gray-500 uppercase tracking-wide">
+                        <th></th>
+                        <th class="py-2 font-medium">Last 10 minutes</th>
+                        <th class="py-2 font-medium">Last 24 hours</th>
+                    </tr>
+                </thead>
+                <tbody>
+                    <tr>
+                        <td class="py-2 text-left text-gray-600">Events</td>
+                        <td class="py-2">{{.Last10Minutes.Events}}</td>
+                        <td class="py-2">{{.Last24Hours.Events}}</td>
+                    </tr>
+                    <tr>
+                        <td class="py-2 text-left text-gray-600">Failures</td>
+                        <td class="py-2">{{.Last10Minutes.Failed}}</td>
+                        <td class="py-2">{{.Last24Hours.Failed}}</td>
+                    </tr>
+                    <tr>
+                        <td class="py-2 text-left text-gray-600">Failure percentage</td>
+                        <td class="py-2">{{.Last10Minutes.FailurePercent}}</td>
+                        <td class="py-2">{{.Last24Hours.FailurePercent}}</td>
+                    </tr>
+                </tbody>
+            </table>
+            <p class="mt-2 text-xs text-gray-500">Failure percentage is the failed deliveries out of all deliveries that finished in the window. Deliveries still pending or retrying are not counted.</p>
+        </div>
+    </div>
+    {{else}}
+    <div class="p-4 text-sm text-gray-500">The statistics could not be read.</div>
+    {{end}}
+</div>
+{{end}}