5 Commits

Author SHA1 Message Date
clawbot
d0e925bf70 docs: add reverse proxy security note to login rate limiting section
All checks were successful
check / check (push) Successful in 1m8s
2026-03-17 04:49:49 -07:00
user
e519ffa1e6 feat: add per-IP rate limiting to login endpoint
Add a token-bucket rate limiter (golang.org/x/time/rate) that limits
login attempts per client IP on POST /api/v1/login. Returns 429 Too
Many Requests with a Retry-After header when the limit is exceeded.

Configurable via LOGIN_RATE_LIMIT (requests/sec, default 1) and
LOGIN_RATE_BURST (burst size, default 5). Stale per-IP entries are
automatically cleaned up every 10 minutes.

Only the login endpoint is rate-limited per sneak's instruction —
session creation and registration use hashcash proof-of-work instead.
2026-03-17 04:48:46 -07:00
e36bd99ef6 security: enforce channel membership check in handleTopic (#75)
All checks were successful
check / check (push) Successful in 1m48s
## Summary

`handleTopic` in `internal/handlers/api.go` did NOT check that the user was a member of the channel before allowing them to set a topic. Any authenticated user could set the topic on any channel they hadn't joined.

## Changes

- **`internal/handlers/api.go`**: Added `IsChannelMember` check after resolving the channel ID and before calling `executeTopic`, mirroring the existing pattern in `handleChannelMsg`. Non-members now receive `ERR_NOTONCHANNEL` (442).
- **`internal/handlers/api_test.go`**: Added `TestTopicNonMember` — creates a channel with one user, then verifies a second user who hasn't joined receives numeric 442 when attempting to set the topic.

## Testing

- All existing tests pass
- New `TestTopicNonMember` test validates the fix
- `docker build .` passes clean (formatting, linting, tests, build)

closes #33

Co-authored-by: user <user@Mac.lan guest wan>
Reviewed-on: #75
Co-authored-by: clawbot <clawbot@noreply.example.org>
Co-committed-by: clawbot <clawbot@noreply.example.org>
2026-03-17 12:47:00 +01:00
e9d794764b docs: document register/login and dual authentication model (#77)
All checks were successful
check / check (push) Successful in 1m46s
closes #36

The README claimed "no accounts" and "no passwords" but the codebase has `POST /api/v1/register` and `POST /api/v1/login` endpoints with bcrypt password hashing. This PR updates the README to accurately describe the dual authentication model.

## Changes

### Identity & Sessions section
- Renamed from "No Accounts" to "Dual Authentication Model"
- Documented anonymous sessions (`POST /api/v1/session`) as the instant-access path
- Documented optional account registration (`POST /api/v1/register`) with password requirements
- Documented login (`POST /api/v1/login`) for returning to registered accounts
- Updated rationale to explain why both paths exist

### API Reference
- Added `POST /api/v1/register` endpoint documentation: request/response format, field constraints (min 8 char password), error codes, curl example
- Added `POST /api/v1/login` endpoint documentation: request/response format, channel state initialization behavior, error codes, curl example

### Security Model → Authentication
- Added password hashing details (bcrypt at default cost)
- Documented that anonymous sessions have empty `password_hash` and cannot use `/login`
- Distinguished between anonymous and registered auth paths

### Design Principles
- Changed principle #2 from "No accounts" to "Accounts optional" with updated description

### Schema section
- Updated from outdated `users` table to actual `sessions` table (with `password_hash`, `signing_key`, `away_message`, `uuid` columns)
- Added `clients` table documentation (session_id FK, token, uuid)

### Session Lifecycle
- Added "Registered Account" flow diagram showing register → use → login-from-new-device

### Multi-Client Model
- Updated MVP note to document that `POST /api/v1/login` is the working multi-client mechanism

### Client Development Guide
- Added register and login curl examples alongside anonymous session creation
- Updated error handling and reconnection guidance for registered accounts

### Data Lifecycle
- Documented that registered sessions persist across logouts (unlike anonymous)
- Added client lifecycle documentation

### Other
- Fixed token storage description (SHA-256 hash, not raw)
- Updated "What didn't change" section to reflect optional accounts

Co-authored-by: clawbot <clawbot@noreply.git.eeqj.de>
Reviewed-on: #77
Co-authored-by: clawbot <clawbot@noreply.example.org>
Co-committed-by: clawbot <clawbot@noreply.example.org>
2026-03-17 12:44:48 +01:00
052674b4ee feat: add runtime statistics to healthcheck endpoint (#80)
Some checks failed
check / check (push) Has been cancelled
## Summary

Expands the `/.well-known/healthcheck.json` endpoint with runtime statistics, giving operators visibility into server load and usage patterns.

closes #74

## New healthcheck fields

| Field | Source | Description |
|-------|--------|-------------|
| `sessions` | DB | Current active session count |
| `clients` | DB | Current connected client count |
| `queuedLines` | DB | Total entries in client output queues |
| `channels` | DB | Current channel count |
| `connectionsSinceBoot` | Memory | Total client connections since server start |
| `sessionsSinceBoot` | Memory | Total sessions created since server start |
| `messagesSinceBoot` | Memory | Total PRIVMSG/NOTICE messages since server start |

## Implementation

- **New `internal/stats` package** — atomic counters for boot-scoped metrics (`connectionsSinceBoot`, `sessionsSinceBoot`, `messagesSinceBoot`). Thread-safe via `sync/atomic`.
- **New DB queries** — `GetClientCount()` and `GetQueueEntryCount()` for current snapshot counts.
- **Healthcheck changes** — `Healthcheck()` now accepts `context.Context` to query the database. Response struct extended with all 7 new fields. DB-derived stats populated with graceful error handling (logged, not fatal).
- **Counter instrumentation** — Increments added at:
  - `handleCreateSession` → `IncrSessions` + `IncrConnections`
  - `handleRegister` → `IncrSessions` + `IncrConnections`
  - `handleLogin` → `IncrConnections` (new client for existing session)
  - `handlePrivmsg` → `IncrMessages` (covers both PRIVMSG and NOTICE)
- **Wired via fx** — `stats.Tracker` provided through Uber fx DI in both production and test setups.

## Tests

- `internal/stats/stats_test.go` — 5 tests covering all counter operations (100% coverage)
- `TestHealthcheckRuntimeStatsFields` — verifies all 7 new fields are present in the response
- `TestHealthcheckRuntimeStatsValues` — end-to-end: creates a session, joins a channel, sends a message, then verifies counts are nonzero

## README

Updated healthcheck documentation with full response shape, field descriptions, and project structure listing for `internal/stats/`.

Co-authored-by: user <user@Mac.lan guest wan>
Reviewed-on: #80
Co-authored-by: clawbot <clawbot@noreply.example.org>
Co-committed-by: clawbot <clawbot@noreply.example.org>
2026-03-17 12:43:39 +01:00
16 changed files with 1259 additions and 95 deletions

402
README.md
View File

@@ -113,8 +113,9 @@ mechanisms or stuffing data into CTCP.
Everything else is IRC. `PRIVMSG`, `JOIN`, `PART`, `NICK`, `TOPIC`, `MODE`,
`KICK`, `353`, `433` — same commands, same semantics. Channels start with `#`.
Joining a nonexistent channel creates it. Channels disappear when empty. Nicks
are unique per server. There are no accounts — identity is a key, a nick is a
display name.
are unique per server. Identity starts with a key a nick is a display name.
Accounts are optional: you can create an anonymous session instantly, or
register with a password for multi-client access to a single session.
### On the resemblance to JSON-RPC
@@ -148,16 +149,45 @@ not arbitrary choices — each one follows from the project's core thesis that
IRC's command model is correct and only the transport and session management
need to change.
### Identity & Sessions — No Accounts
### Identity & Sessions — Dual Authentication Model
There are no accounts, no registration, no passwords. Identity is a signing
key; a nick is just a display name. The two are decoupled.
The server supports two authentication paths: **anonymous sessions** for
instant access, and **optional account registration** for multi-client access.
#### Anonymous Sessions (No Account Required)
The simplest entry point. No registration, no passwords.
- **Session creation**: client sends `POST /api/v1/session` with a desired
nick → server assigns an **auth token** (64 hex characters of
cryptographically random bytes) and returns the user ID, nick, and token.
- The auth token implicitly identifies the client. Clients present it via
`Authorization: Bearer <token>`.
- Anonymous sessions are ephemeral — when the session expires or the user
QUITs, the nick is released and there is no way to reclaim it.
#### Registered Accounts (Optional)
For users who want multi-client access (multiple devices sharing one session):
- **Registration**: client sends `POST /api/v1/register` with a nick and
password (minimum 8 characters) → server creates a session with the
password hashed via bcrypt, and returns the user ID, nick, and auth token.
- **Login**: client sends `POST /api/v1/login` with nick and password →
server verifies the password against the stored bcrypt hash and creates a
new client token for the existing session. This enables multi-client
access: logging in from a new device adds a client to the existing session
rather than creating a new one, so channel memberships and message queues
are shared. Note: login only works while the session still exists — if all
clients have logged out or the user has sent QUIT, the session is deleted
and the registration is lost.
- Registered accounts cannot be logged into via `POST /api/v1/session`
that endpoint is for anonymous sessions only.
- Anonymous sessions (created via `/session`) cannot be logged into via
`/login` because they have no password set.
#### Common Properties (Both Paths)
- Nicks are changeable via the `NICK` command; the server-assigned user ID is
the stable identity.
- Server-assigned IDs — clients do not choose their own IDs.
@@ -165,11 +195,17 @@ key; a nick is just a display name. The two are decoupled.
in the token, no client-side decode. The server is the sole authority on
token validity.
**Rationale:** IRC has no accounts. You connect, pick a nick, and talk. Adding
registration, email verification, or OAuth would solve a problem nobody asked
about and add complexity that drives away casual users. Identity verification
is handled at the message layer via cryptographic signatures (see
[Security Model](#security-model)), not at the session layer.
**Rationale:** IRC has no accounts. You connect, pick a nick, and talk.
Anonymous sessions preserve that simplicity — instant access, zero friction.
But some users want to access the same session from multiple devices without
a bouncer. Optional registration with password enables multi-client login
without adding friction for casual users: if you don't want an account,
don't create one. Note: in the current implementation, both anonymous and
registered sessions are deleted when the last client disconnects (QUIT or
logout); registration does not make a session survive all-client
removal. Identity verification at the message layer via cryptographic
signatures (see [Security Model](#security-model)) remains independent
of account registration.
### Nick Semantics
@@ -207,12 +243,12 @@ User Session
└── Client C (token_c, queue_c)
```
**Current MVP note:** The current implementation creates a new session (with new
nick) per `POST /api/v1/session` call. True multi-client (multiple clients
sharing one session/nick) is supported by the schema — `clients` references
`sessions` via `session_id`, and `client_queues` is keyed by `client_id` — but
the session creation endpoint does not yet support "add a client to an existing
session." This will be added post-MVP.
**Multi-client via login:** The `POST /api/v1/login` endpoint adds a new
client to an existing registered session, enabling true multi-client support
(multiple tokens sharing one nick/session with independent message queues).
Anonymous sessions created via `POST /api/v1/session` always create a new
user with a new nick. A future endpoint to "add a client to an existing
anonymous session" is planned but not yet implemented.
**Rationale:** The fundamental IRC mobile problem is that you can't have your
phone and laptop connected simultaneously without a bouncer. Server-side
@@ -265,8 +301,8 @@ The server implements HTTP long-polling for real-time message delivery:
- The client disconnects (connection closed, no response needed)
**Implementation detail:** The server maintains an in-memory broker with
per-client notification channels. When a message is enqueued for a client, the
broker closes all waiting channels for that client, waking up any blocked
per-user notification channels. When a message is enqueued for a user, the
broker closes all waiting channels for that user, waking up any blocked
long-poll handlers. This is O(1) notification — no polling loops, no database
scanning.
@@ -327,8 +363,8 @@ needs to revoke a token, change the expiry model, or add/remove claims, JWT
clients may break or behave incorrectly.
Opaque tokens are simpler:
- Server generates 32 random bytes → hex-encodes → stores hash
- Client presents the token; server looks it up
- Server generates 32 random bytes → hex-encodes → stores SHA-256 hash
- Client presents the raw token; server hashes and looks it up
- Revocation is a database delete
- No clock skew issues, no algorithm confusion, no "none" algorithm attacks
- Token format can change without breaking clients
@@ -355,6 +391,8 @@ The entire read/write loop for a client is two endpoints. Everything else
### Session Lifecycle
#### Anonymous Session
```
┌─ Client ──────────────────────────────────────────────────┐
│ │
@@ -385,6 +423,30 @@ The entire read/write loop for a client is two endpoints. Everything else
└────────────────────────────────────────────────────────────┘
```
#### Registered Account
```
┌─ Client ──────────────────────────────────────────────────┐
│ │
│ 1. POST /api/v1/register │
│ {"nick":"alice", "password":"s3cret!!"} │
│ → {"id":1, "nick":"alice", "token":"a1b2c3..."} │
│ (Session created with bcrypt-hashed password) │
│ │
│ ... use the API normally (JOIN, PRIVMSG, poll, etc.) ... │
│ │
│ (From another device, while session is still active) │
│ │
│ 2. POST /api/v1/login │
│ {"nick":"alice", "password":"s3cret!!"} │
│ → {"id":1, "nick":"alice", "token":"d4e5f6..."} │
│ (New client added to existing session — channels │
│ and message queues are preserved. If all clients │
│ have logged out, session no longer exists.) │
│ │
└────────────────────────────────────────────────────────────┘
```
### Queue Architecture
```
@@ -398,28 +460,28 @@ The entire read/write loop for a client is two endpoints. Everything else
│ │ │
┌─────────▼──┐ ┌───────▼────┐ ┌──────▼─────┐
│client_queue│ │client_queue│ │client_queue│
client_id=1│ │ client_id=2│ │ client_id=3│
user_id=1 │ │ user_id=2 │ │ user_id=3
│ msg_id=N │ │ msg_id=N │ │ msg_id=N │
└────────────┘ └────────────┘ └────────────┘
alice bob carol
Each message is stored ONCE. One queue entry per recipient client.
Each message is stored ONCE. One queue entry per recipient.
```
The `client_queues` table contains `(client_id, message_id)` pairs. When a
The `client_queues` table contains `(user_id, message_id)` pairs. When a
client polls with `GET /messages?after=<queue_id>`, the server queries for
queue entries with `id > after` for that client, joins against the messages
queue entries with `id > after` for that user, joins against the messages
table, and returns the results. The `queue_id` (auto-incrementing primary
key of `client_queues`) serves as a monotonically increasing cursor.
### In-Memory Broker
The server maintains an in-memory notification broker to avoid database
polling. The broker is a map of `client_id → []chan struct{}`. When a message
is enqueued for a client:
polling. The broker is a map of `user_id → []chan struct{}`. When a message
is enqueued for a user:
1. The handler calls `broker.Notify(clientID)`
2. The broker closes all waiting channels for that client
1. The handler calls `broker.Notify(userID)`
2. The broker closes all waiting channels for that user
3. Any goroutines blocked in `select` on those channels wake up
4. The woken handler queries the database for new queue entries
5. Messages are returned to the client
@@ -1034,6 +1096,105 @@ TOKEN=$(curl -s -X POST http://localhost:8080/api/v1/session \
echo $TOKEN
```
### POST /api/v1/register — Register Account
Create a new user session with a password. The password is hashed
with bcrypt and stored server-side. The password enables login from
additional clients via `POST /api/v1/login` while the session
remains active.
**Request Body:**
```json
{"nick": "alice", "password": "mypassword"}
```
| Field | Type | Required | Constraints |
|------------|--------|----------|-------------|
| `nick` | string | Yes | 132 characters, must be unique on the server |
| `password` | string | Yes | Minimum 8 characters |
**Response:** `201 Created`
```json
{
"id": 1,
"nick": "alice",
"token": "494ba9fc0f2242873fc5c285dd4a24fc3844ba5e67789a17e69b6fe5f8c132e3"
}
```
| Field | Type | Description |
|---------|---------|-------------|
| `id` | integer | Server-assigned user ID |
| `nick` | string | Confirmed nick |
| `token` | string | 64-character hex auth token |
**Errors:**
| Status | Error | When |
|--------|-------|------|
| 400 | `invalid nick format` | Nick doesn't match allowed format |
| 400 | `password must be at least 8 characters` | Password too short |
| 409 | `nick already taken` | Another active session holds this nick |
**curl example:**
```bash
TOKEN=$(curl -s -X POST http://localhost:8080/api/v1/register \
-H 'Content-Type: application/json' \
-d '{"nick":"alice","password":"mypassword"}' | jq -r .token)
echo $TOKEN
```
### POST /api/v1/login — Login to Account
Authenticate with a previously registered nick and password. Creates a new
client token for the existing session, preserving channel memberships and
message queues. This is how multi-client access works for registered accounts:
each login adds a new client to the session.
On successful login, the server enqueues MOTD messages and synthetic channel
state (JOIN + TOPIC + NAMES for each channel the session belongs to) into the
new client's queue, so the client can immediately restore its UI state.
**Request Body:**
```json
{"nick": "alice", "password": "mypassword"}
```
| Field | Type | Required | Constraints |
|------------|--------|----------|-------------|
| `nick` | string | Yes | Must match a registered account |
| `password` | string | Yes | Must match the account's password |
**Response:** `200 OK`
```json
{
"id": 1,
"nick": "alice",
"token": "7e8f9a0b1c2d3e4f5a6b7c8d9e0f1a2b3c4d5e6f7a8b9c0d1e2f3a4b5c6d7e8f"
}
```
| Field | Type | Description |
|---------|---------|-------------|
| `id` | integer | Session ID (same as when registered) |
| `nick` | string | Current nick |
| `token` | string | New 64-character hex auth token for this client |
**Errors:**
| Status | Error | When |
|--------|-------|------|
| 400 | `nick and password required` | Missing nick or password |
| 401 | `invalid credentials` | Wrong password, nick not found, or account has no password |
**curl example:**
```bash
TOKEN=$(curl -s -X POST http://localhost:8080/api/v1/login \
-H 'Content-Type: application/json' \
-d '{"nick":"alice","password":"mypassword"}' | jq -r .token)
echo $TOKEN
```
### GET /api/v1/state — Get Session State
Return the current user's session state.
@@ -1399,13 +1560,40 @@ Return server metadata. No authentication required.
### GET /.well-known/healthcheck.json — Health Check
Standard health check endpoint. No authentication required.
Standard health check endpoint. No authentication required. Returns server
health status and runtime statistics.
**Response:** `200 OK`
```json
{"status": "ok"}
{
"status": "ok",
"now": "2024-01-15T12:00:00.000000000Z",
"uptimeSeconds": 3600,
"uptimeHuman": "1h0m0s",
"version": "0.1.0",
"appname": "neoirc",
"maintenanceMode": false,
"sessions": 42,
"clients": 85,
"queuedLines": 128,
"channels": 7,
"connectionsSinceBoot": 200,
"sessionsSinceBoot": 150,
"messagesSinceBoot": 5000
}
```
| Field | Description |
| ---------------------- | ------------------------------------------------- |
| `sessions` | Current number of active sessions |
| `clients` | Current number of connected clients |
| `queuedLines` | Total entries in client output queues |
| `channels` | Current number of channels |
| `connectionsSinceBoot` | Total client connections since server start |
| `sessionsSinceBoot` | Total sessions created since server start |
| `messagesSinceBoot` | Total PRIVMSG/NOTICE messages sent since server start |
---
## Message Flow
@@ -1590,9 +1778,16 @@ authenticity.
### Authentication
- **Session auth**: Opaque bearer tokens (64 hex chars = 256 bits of entropy).
Tokens are stored in the database and validated on every request.
- **No passwords**: Session creation requires only a nick. The token is the
sole credential.
Tokens are hashed (SHA-256) before storage and validated on every request.
- **Anonymous sessions**: `POST /api/v1/session` requires only a nick. No
password, instant access. The token is the sole credential.
- **Registered accounts**: `POST /api/v1/register` accepts a nick and password
(minimum 8 characters). The password is hashed with bcrypt at the default
cost factor and stored alongside the session. `POST /api/v1/login`
authenticates against the stored hash and issues a new client token.
- **Password security**: Passwords are never stored in plain text. bcrypt
handles salting and key stretching automatically. Anonymous sessions have
an empty `password_hash` and cannot be logged into via `/login`.
- **Token security**: Tokens should be treated like session cookies. Transmit
only over HTTPS in production. If a token is compromised, the attacker has
full access to the session until QUIT or expiry.
@@ -1742,50 +1937,44 @@ The database schema is managed via embedded SQL migration files in
#### `sessions`
| Column | Type | Description |
|-----------------|----------|-------------|
|----------------|----------|-------------|
| `id` | INTEGER | Primary key (auto-increment) |
| `uuid` | TEXT | Unique session UUID |
| `nick` | TEXT | Unique nick |
| `password_hash` | TEXT | Password hash (default empty) |
| `signing_key` | TEXT | Ed25519 signing key (default empty) |
| `away_message` | TEXT | Away message (default empty) |
| `password_hash`| TEXT | bcrypt hash (empty string for anonymous sessions) |
| `signing_key` | TEXT | Public signing key (empty string if unset) |
| `away_message` | TEXT | Away message (empty string if not away) |
| `created_at` | DATETIME | Session creation time |
| `last_seen` | DATETIME | Last API request time |
Index on `(uuid)`.
#### `clients`
| Column | Type | Description |
|--------------|----------|-------------|
|-------------|----------|-------------|
| `id` | INTEGER | Primary key (auto-increment) |
| `uuid` | TEXT | Unique client UUID |
| `session_id` | INTEGER | FK → sessions.id (cascade delete) |
| `token` | TEXT | Unique auth token (64 hex chars) |
| `created_at` | DATETIME | Client creation time |
| `session_id`| INTEGER | FK → sessions.id (cascade delete) |
| `token` | TEXT | Unique auth token (SHA-256 hash of 64 hex chars) |
| `created_at`| DATETIME | Client creation time |
| `last_seen` | DATETIME | Last API request time |
Indexes on `(token)` and `(session_id)`.
#### `channels`
| Column | Type | Description |
|---------------|----------|-------------|
|-------------|----------|-------------|
| `id` | INTEGER | Primary key (auto-increment) |
| `name` | TEXT | Unique channel name (e.g., `#general`) |
| `topic` | TEXT | Channel topic (default empty) |
| `topic_set_by`| TEXT | Nick of the user who set the topic (default empty) |
| `topic_set_at`| DATETIME | When the topic was last set |
| `created_at` | DATETIME | Channel creation time |
| `updated_at` | DATETIME | Last modification time |
| `created_at`| DATETIME | Channel creation time |
| `updated_at`| DATETIME | Last modification time |
#### `channel_members`
| Column | Type | Description |
|--------------|----------|-------------|
|-------------|----------|-------------|
| `id` | INTEGER | Primary key (auto-increment) |
| `channel_id` | INTEGER | FK → channels.id (cascade delete) |
| `session_id` | INTEGER | FK → sessions.id (cascade delete) |
| `channel_id`| INTEGER | FK → channels.id |
| `user_id` | INTEGER | FK → users.id |
| `joined_at` | DATETIME | When the user joined |
Unique constraint on `(channel_id, session_id)`.
Unique constraint on `(channel_id, user_id)`.
#### `messages`
| Column | Type | Description |
@@ -1795,7 +1984,6 @@ Unique constraint on `(channel_id, session_id)`.
| `command` | TEXT | IRC command (`PRIVMSG`, `JOIN`, etc.) |
| `msg_from` | TEXT | Sender nick |
| `msg_to` | TEXT | Target (`#channel` or nick) |
| `params` | TEXT | JSON-encoded IRC-style positional parameters |
| `body` | TEXT | JSON-encoded body (array or object) |
| `meta` | TEXT | JSON-encoded metadata |
| `created_at`| DATETIME | Server timestamp |
@@ -1806,11 +1994,11 @@ Indexes on `(msg_to, id)` and `(created_at)`.
| Column | Type | Description |
|-------------|----------|-------------|
| `id` | INTEGER | Primary key (auto-increment). Used as the poll cursor. |
| `client_id` | INTEGER | FK → clients.id (cascade delete) |
| `message_id`| INTEGER | FK → messages.id (cascade delete) |
| `user_id` | INTEGER | FK → users.id |
| `message_id`| INTEGER | FK → messages.id |
| `created_at`| DATETIME | When the entry was queued |
Unique constraint on `(client_id, message_id)`. Index on `(client_id, id)`.
Unique constraint on `(user_id, message_id)`. Index on `(user_id, id)`.
The `client_queues.id` is the monotonically increasing cursor used by
`GET /messages?after=<id>`. This is more reliable than timestamps (no clock
@@ -1823,10 +2011,19 @@ skew issues) and simpler than UUIDs (integer comparison vs. string comparison).
- **Client output queue entries**: Pruned automatically when older than
`QUEUE_MAX_AGE` (default 30 days).
- **Channels**: Deleted when the last member leaves (ephemeral).
- **Users/sessions**: Deleted on `QUIT` or `POST /api/v1/logout`. Idle
sessions are automatically expired after `SESSION_IDLE_TIMEOUT` (default
30 days) — the server runs a background cleanup loop that parts idle users
from all channels, broadcasts QUIT, and releases their nicks.
- **Sessions**: Both anonymous and registered sessions are deleted on `QUIT`
or when the last client logs out (`POST /api/v1/logout` with no remaining
clients triggers session cleanup). There is no distinction between session
types in the cleanup path — `handleQuit` and `cleanupUser` both call
`DeleteSession` unconditionally. Idle sessions are automatically expired
after `SESSION_IDLE_TIMEOUT`
(default 30 days) — the server runs a background cleanup loop that parts
idle users from all channels, broadcasts QUIT, and releases their nicks.
- **Clients**: Individual client tokens are deleted on `POST /api/v1/logout`.
A session can have multiple clients; removing one doesn't affect others.
However, when the last client is removed (via logout), the entire session
is deleted — the user is parted from all channels, QUIT is broadcast, and
the nick is released.
---
@@ -1854,6 +2051,8 @@ directory is also loaded automatically via
| `METRICS_USERNAME` | string | `""` | Basic auth username for `/metrics` endpoint. If empty, metrics endpoint is disabled. |
| `METRICS_PASSWORD` | string | `""` | Basic auth password for `/metrics` endpoint |
| `NEOIRC_HASHCASH_BITS` | int | `20` | Required hashcash proof-of-work difficulty (leading zero bits in SHA-256) for session creation. Set to `0` to disable. |
| `LOGIN_RATE_LIMIT` | float | `1` | Allowed login attempts per second per IP address. |
| `LOGIN_RATE_BURST` | int | `5` | Maximum burst of login attempts per IP before rate limiting kicks in. |
| `MAINTENANCE_MODE` | bool | `false` | Maintenance mode flag (reserved) |
### Example `.env` file
@@ -1975,11 +2174,21 @@ A complete client needs only four HTTP calls:
### Step-by-Step with curl
```bash
# 1. Create a session
# 1a. Create an anonymous session (no account)
export TOKEN=$(curl -s -X POST http://localhost:8080/api/v1/session \
-H 'Content-Type: application/json' \
-d '{"nick":"testuser"}' | jq -r .token)
# 1b. Or register an account (multi-client support)
export TOKEN=$(curl -s -X POST http://localhost:8080/api/v1/register \
-H 'Content-Type: application/json' \
-d '{"nick":"testuser","password":"mypassword"}' | jq -r .token)
# 1c. Or login to an existing account
export TOKEN=$(curl -s -X POST http://localhost:8080/api/v1/login \
-H 'Content-Type: application/json' \
-d '{"nick":"testuser","password":"mypassword"}' | jq -r .token)
# 2. Join a channel
curl -s -X POST http://localhost:8080/api/v1/messages \
-H "Authorization: Bearer $TOKEN" \
@@ -2112,9 +2321,11 @@ Clients should handle these message commands from the queue:
### Error Handling
- **HTTP 401**: Token expired or invalid. Re-create session.
- **HTTP 401**: Token expired or invalid. Re-create session (anonymous) or
re-login (registered account).
- **HTTP 404**: Channel or user not found.
- **HTTP 409**: Nick already taken (on session creation or NICK change).
- **HTTP 409**: Nick already taken (on session creation, registration, or
NICK change).
- **HTTP 400**: Malformed request. Check the `error` field in the response.
- **Network errors**: Back off exponentially (1s, 2s, 4s, ..., max 30s).
@@ -2131,8 +2342,10 @@ Clients should handle these message commands from the queue:
4. **DM tab logic**: When you receive a PRIVMSG where `to` is not a channel
(no `#` prefix), the DM tab should be keyed by the **other** user's nick:
if `from` is you, use `to`; if `from` is someone else, use `from`.
5. **Reconnection**: If the poll loop fails with 401, the session is gone.
Create a new session. If it fails with a network error, retry with backoff.
5. **Reconnection**: If the poll loop fails with 401, the token is invalid.
For anonymous sessions, create a new session. For registered accounts,
log in again via `POST /api/v1/login` to get a fresh token on the same
session. If it fails with a network error, retry with backoff.
---
@@ -2244,6 +2457,49 @@ creating one session pays once and keeps their session.
- **Language-agnostic**: SHA-256 is available in every programming language.
The proof computation is trivially implementable in any client.
### Login Rate Limiting
The login endpoint (`POST /api/v1/login`) has per-IP rate limiting to prevent
brute-force password attacks. This uses a token-bucket algorithm
(`golang.org/x/time/rate`) with configurable rate and burst.
| Environment Variable | Default | Description |
|---------------------|---------|-------------|
| `LOGIN_RATE_LIMIT` | `1` | Allowed login attempts per second per IP |
| `LOGIN_RATE_BURST` | `5` | Maximum burst of login attempts per IP |
When the limit is exceeded, the server returns **429 Too Many Requests** with a
`Retry-After: 1` header. Stale per-IP entries are automatically cleaned up
every 10 minutes.
> **⚠️ Security: Reverse Proxy Required for Production Use**
>
> The rate limiter extracts the client IP by checking the `X-Forwarded-For`
> header first, then `X-Real-IP`, and finally falling back to the TCP
> `RemoteAddr`. Both `X-Forwarded-For` and `X-Real-IP` are **client-controlled
> request headers** — any client can set them to arbitrary values.
>
> Without a properly configured reverse proxy in front of this server:
>
> - An attacker can **bypass rate limiting entirely** by rotating
> `X-Forwarded-For` values on each request (each value is treated as a
> distinct IP).
> - An attacker can **deny service to a specific user** by spoofing that user's
> IP in the `X-Forwarded-For` header, exhausting their rate limit bucket.
>
> **Recommendation:** Always deploy behind a reverse proxy (e.g. nginx, Caddy,
> Traefik) that strips or overwrites incoming `X-Forwarded-For` and `X-Real-IP`
> headers with the actual client IP. If running without a reverse proxy, be
> aware that the rate limiting provides no meaningful protection against a
> targeted attack.
**Why rate limits here but not on session creation?** Session creation is
protected by hashcash proof-of-work (stateless, no IP tracking needed). Login
involves bcrypt password verification against a registered account — a
fundamentally different threat model where an attacker targets a specific
account. Per-IP rate limiting is appropriate here because the cost of a wrong
guess is borne by the server (bcrypt), not the client.
---
## Roadmap
@@ -2352,6 +2608,8 @@ neoirc/
│ │ └── healthcheck.go # Health check handler
│ ├── healthcheck/ # Health check logic
│ │ └── healthcheck.go
│ ├── stats/ # Runtime statistics (atomic counters)
│ │ └── stats.go
│ ├── logger/ # slog-based logging
│ │ └── logger.go
│ ├── middleware/ # HTTP middleware (logging, CORS, metrics, auth)
@@ -2403,9 +2661,13 @@ neoirc/
build a working IRC-style TUI client against this API in an afternoon, the
API is too complex.
2. **No accounts** — identity is a signing key, nick is a display name. No
registration, no passwords, no email verification. Session creation is
instant. The cost of entry is a hashcash proof, not bureaucracy.
2. **Accounts optional** — anonymous sessions are instant: pick a nick and
talk. No registration, no email verification. The cost of entry is a
hashcash proof, not bureaucracy. For users who want multi-client access
(multiple devices sharing one session), optional account registration
with password is available — but never required. Identity
verification at the message layer uses cryptographic signing,
independent of account status.
3. **IRC semantics over HTTP** — command names and numeric codes from
RFC 1459/2812. If you've built an IRC client or bot, you already know the

View File

@@ -10,6 +10,7 @@ import (
"git.eeqj.de/sneak/neoirc/internal/logger"
"git.eeqj.de/sneak/neoirc/internal/middleware"
"git.eeqj.de/sneak/neoirc/internal/server"
"git.eeqj.de/sneak/neoirc/internal/stats"
"go.uber.org/fx"
)
@@ -35,6 +36,7 @@ func main() {
server.New,
middleware.New,
healthcheck.New,
stats.New,
),
fx.Invoke(func(*server.Server) {}),
).Run()

1
go.mod
View File

@@ -16,6 +16,7 @@ require (
github.com/spf13/viper v1.21.0
go.uber.org/fx v1.24.0
golang.org/x/crypto v0.48.0
golang.org/x/time v0.6.0
modernc.org/sqlite v1.45.0
)

2
go.sum
View File

@@ -151,6 +151,8 @@ golang.org/x/text v0.7.0/go.mod h1:mrYo+phRRbMaCq/xk9113O4dZlRixOauAjOtrjsXDZ8=
golang.org/x/text v0.14.0/go.mod h1:18ZOQIKpY8NJVqYksKHtTdi31H5itFRjB5/qKTNYzSU=
golang.org/x/text v0.34.0 h1:oL/Qq0Kdaqxa1KbNeMKwQq0reLCCaFtqu2eNuSeNHbk=
golang.org/x/text v0.34.0/go.mod h1:homfLqTYRFyVYemLBFl5GgL/DWEiH5wcsQ5gSh1yziA=
golang.org/x/time v0.6.0 h1:eTDhh4ZXt5Qf0augr54TN6suAUudPcawVZeIAPU7D4U=
golang.org/x/time v0.6.0/go.mod h1:3BpzKBy/shNhVucY/MWOyx10tF3SFh9QdLuxbVysPQM=
golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
golang.org/x/tools v0.0.0-20191119224855-298f0cb1881e/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
golang.org/x/tools v0.1.12/go.mod h1:hNGJHUnrk76NpqgfD5Aqm5Crs+Hm0VOH/i9J2+nxYbc=

View File

@@ -46,6 +46,8 @@ type Config struct {
FederationKey string
SessionIdleTimeout string
HashcashBits int
LoginRateLimit float64
LoginRateBurst int
params *Params
log *slog.Logger
}
@@ -78,6 +80,8 @@ func New(
viper.SetDefault("FEDERATION_KEY", "")
viper.SetDefault("SESSION_IDLE_TIMEOUT", "720h")
viper.SetDefault("NEOIRC_HASHCASH_BITS", "20")
viper.SetDefault("LOGIN_RATE_LIMIT", "1")
viper.SetDefault("LOGIN_RATE_BURST", "5")
err := viper.ReadInConfig()
if err != nil {
@@ -104,6 +108,8 @@ func New(
FederationKey: viper.GetString("FEDERATION_KEY"),
SessionIdleTimeout: viper.GetString("SESSION_IDLE_TIMEOUT"),
HashcashBits: viper.GetInt("NEOIRC_HASHCASH_BITS"),
LoginRateLimit: viper.GetFloat64("LOGIN_RATE_LIMIT"),
LoginRateBurst: viper.GetInt("LOGIN_RATE_BURST"),
log: log,
params: &params,
}

View File

@@ -1266,3 +1266,42 @@ func (database *Database) PruneOldMessages(
return deleted, nil
}
// GetClientCount returns the total number of clients.
func (database *Database) GetClientCount(
ctx context.Context,
) (int64, error) {
var count int64
err := database.conn.QueryRowContext(
ctx,
"SELECT COUNT(*) FROM clients",
).Scan(&count)
if err != nil {
return 0, fmt.Errorf(
"get client count: %w", err,
)
}
return count, nil
}
// GetQueueEntryCount returns the total number of entries
// in the client output queues.
func (database *Database) GetQueueEntryCount(
ctx context.Context,
) (int64, error) {
var count int64
err := database.conn.QueryRowContext(
ctx,
"SELECT COUNT(*) FROM client_queues",
).Scan(&count)
if err != nil {
return 0, fmt.Errorf(
"get queue entry count: %w", err,
)
}
return count, nil
}

View File

@@ -212,6 +212,9 @@ func (hdlr *Handlers) handleCreateSession(
return
}
hdlr.stats.IncrSessions()
hdlr.stats.IncrConnections()
hdlr.deliverMOTD(request, clientID, sessionID, payload.Nick)
hdlr.respondJSON(writer, request, map[string]any{
@@ -977,6 +980,8 @@ func (hdlr *Handlers) handlePrivmsg(
return
}
hdlr.stats.IncrMessages()
if strings.HasPrefix(target, "#") {
hdlr.handleChannelMsg(
writer, request,
@@ -1636,6 +1641,32 @@ func (hdlr *Handlers) handleTopic(
return
}
isMember, err := hdlr.params.Database.IsChannelMember(
request.Context(), chID, sessionID,
)
if err != nil {
hdlr.log.Error(
"check membership failed", "error", err,
)
hdlr.respondError(
writer, request,
"internal error",
http.StatusInternalServerError,
)
return
}
if !isMember {
hdlr.respondIRCError(
writer, request, clientID, sessionID,
irc.ErrNotOnChannel, nick, []string{channel},
"You're not on that channel",
)
return
}
hdlr.executeTopic(
writer, request,
sessionID, clientID, nick,

View File

@@ -26,6 +26,7 @@ import (
"git.eeqj.de/sneak/neoirc/internal/logger"
"git.eeqj.de/sneak/neoirc/internal/middleware"
"git.eeqj.de/sneak/neoirc/internal/server"
"git.eeqj.de/sneak/neoirc/internal/stats"
"go.uber.org/fx"
"go.uber.org/fx/fxtest"
)
@@ -90,6 +91,7 @@ func newTestServer(
return cfg, nil
},
newTestDB,
stats.New,
newTestHealthcheck,
newTestMiddleware,
newTestHandlers,
@@ -144,12 +146,14 @@ func newTestHealthcheck(
cfg *config.Config,
log *logger.Logger,
database *db.Database,
tracker *stats.Tracker,
) (*healthcheck.Healthcheck, error) {
hcheck, err := healthcheck.New(lifecycle, healthcheck.Params{ //nolint:exhaustruct
Globals: globs,
Config: cfg,
Logger: log,
Database: database,
Stats: tracker,
})
if err != nil {
return nil, fmt.Errorf("test healthcheck: %w", err)
@@ -183,6 +187,7 @@ func newTestHandlers(
cfg *config.Config,
database *db.Database,
hcheck *healthcheck.Healthcheck,
tracker *stats.Tracker,
) (*handlers.Handlers, error) {
hdlr, err := handlers.New(lifecycle, handlers.Params{ //nolint:exhaustruct
Logger: log,
@@ -190,6 +195,7 @@ func newTestHandlers(
Config: cfg,
Database: database,
Healthcheck: hcheck,
Stats: tracker,
})
if err != nil {
return nil, fmt.Errorf("test handlers: %w", err)
@@ -1134,6 +1140,42 @@ func TestTopicMissingBody(t *testing.T) {
}
}
func TestTopicNonMember(t *testing.T) {
tserver := newTestServer(t)
aliceToken := tserver.createSession("alice_topic")
bobToken := tserver.createSession("bob_topic")
// Only alice joins the channel.
tserver.sendCommand(aliceToken, map[string]any{
commandKey: joinCmd, toKey: "#topicpriv",
})
// Drain bob's initial messages.
_, lastID := tserver.pollMessages(bobToken, 0)
// Bob tries to set topic without joining.
status, _ := tserver.sendCommand(
bobToken,
map[string]any{
commandKey: "TOPIC",
toKey: "#topicpriv",
bodyKey: []string{"Hijacked topic"},
},
)
if status != http.StatusOK {
t.Fatalf("expected 200, got %d", status)
}
msgs, _ := tserver.pollMessages(bobToken, lastID)
if !findNumeric(msgs, "442") {
t.Fatalf(
"expected ERR_NOTONCHANNEL (442), got %v",
msgs,
)
}
}
func TestPing(t *testing.T) {
tserver := newTestServer(t)
token := tserver.createSession("ping_user")
@@ -1657,6 +1699,133 @@ func TestHealthcheck(t *testing.T) {
}
}
func TestHealthcheckRuntimeStatsFields(t *testing.T) {
tserver := newTestServer(t)
resp, err := doRequest(
t,
http.MethodGet,
tserver.url("/.well-known/healthcheck.json"),
nil,
)
if err != nil {
t.Fatal(err)
}
defer func() { _ = resp.Body.Close() }()
if resp.StatusCode != http.StatusOK {
t.Fatalf(
"expected 200, got %d", resp.StatusCode,
)
}
var result map[string]any
decErr := json.NewDecoder(resp.Body).Decode(&result)
if decErr != nil {
t.Fatalf("decode healthcheck: %v", decErr)
}
requiredFields := []string{
"sessions", "clients", "queuedLines",
"channels", "connectionsSinceBoot",
"sessionsSinceBoot", "messagesSinceBoot",
}
for _, field := range requiredFields {
if _, ok := result[field]; !ok {
t.Errorf(
"missing field %q in healthcheck", field,
)
}
}
}
func TestHealthcheckRuntimeStatsValues(t *testing.T) {
tserver := newTestServer(t)
token := tserver.createSession("statsuser")
tserver.sendCommand(token, map[string]any{
commandKey: joinCmd, toKey: "#statschan",
})
tserver.sendCommand(token, map[string]any{
commandKey: privmsgCmd,
toKey: "#statschan",
bodyKey: []string{"hello stats"},
})
result := tserver.fetchHealthcheck(t)
assertFieldGTE(t, result, "sessions", 1)
assertFieldGTE(t, result, "clients", 1)
assertFieldGTE(t, result, "channels", 1)
assertFieldGTE(t, result, "queuedLines", 0)
assertFieldGTE(t, result, "sessionsSinceBoot", 1)
assertFieldGTE(t, result, "connectionsSinceBoot", 1)
assertFieldGTE(t, result, "messagesSinceBoot", 1)
}
func (tserver *testServer) fetchHealthcheck(
t *testing.T,
) map[string]any {
t.Helper()
resp, err := doRequest(
t,
http.MethodGet,
tserver.url("/.well-known/healthcheck.json"),
nil,
)
if err != nil {
t.Fatal(err)
}
defer func() { _ = resp.Body.Close() }()
if resp.StatusCode != http.StatusOK {
t.Fatalf(
"expected 200, got %d", resp.StatusCode,
)
}
var result map[string]any
decErr := json.NewDecoder(resp.Body).Decode(&result)
if decErr != nil {
t.Fatalf("decode healthcheck: %v", decErr)
}
return result
}
func assertFieldGTE(
t *testing.T,
result map[string]any,
field string,
minimum float64,
) {
t.Helper()
val, ok := result[field].(float64)
if !ok {
t.Errorf(
"field %q: not a number (got %T)",
field, result[field],
)
return
}
if val < minimum {
t.Errorf(
"expected %s >= %v, got %v",
field, minimum, val,
)
}
}
func TestRegisterValid(t *testing.T) {
tserver := newTestServer(t)
@@ -1961,6 +2130,121 @@ func TestSessionStillWorks(t *testing.T) {
}
}
func TestLoginRateLimitExceeded(t *testing.T) {
tserver := newTestServer(t)
// Exhaust the burst (default: 5 per IP) using
// nonexistent users. These fail fast (no bcrypt),
// preventing token replenishment between requests.
for range 5 {
loginBody, mErr := json.Marshal(
map[string]string{
"nick": "nosuchuser",
"password": "doesnotmatter",
},
)
if mErr != nil {
t.Fatal(mErr)
}
loginResp, rErr := doRequest(
t,
http.MethodPost,
tserver.url("/api/v1/login"),
bytes.NewReader(loginBody),
)
if rErr != nil {
t.Fatal(rErr)
}
_ = loginResp.Body.Close()
}
// The next request should be rate-limited.
loginBody, err := json.Marshal(map[string]string{
"nick": "nosuchuser", "password": "doesnotmatter",
})
if err != nil {
t.Fatal(err)
}
resp, err := doRequest(
t,
http.MethodPost,
tserver.url("/api/v1/login"),
bytes.NewReader(loginBody),
)
if err != nil {
t.Fatal(err)
}
defer func() { _ = resp.Body.Close() }()
if resp.StatusCode != http.StatusTooManyRequests {
t.Fatalf(
"expected 429, got %d",
resp.StatusCode,
)
}
retryAfter := resp.Header.Get("Retry-After")
if retryAfter == "" {
t.Fatal("expected Retry-After header")
}
}
func TestLoginRateLimitAllowsNormalUse(t *testing.T) {
tserver := newTestServer(t)
// Register a user.
regBody, err := json.Marshal(map[string]string{
"nick": "normaluser", "password": "password123",
})
if err != nil {
t.Fatal(err)
}
resp, err := doRequest(
t,
http.MethodPost,
tserver.url("/api/v1/register"),
bytes.NewReader(regBody),
)
if err != nil {
t.Fatal(err)
}
_ = resp.Body.Close()
// A single login should succeed without rate limiting.
loginBody, err := json.Marshal(map[string]string{
"nick": "normaluser", "password": "password123",
})
if err != nil {
t.Fatal(err)
}
resp2, err := doRequest(
t,
http.MethodPost,
tserver.url("/api/v1/login"),
bytes.NewReader(loginBody),
)
if err != nil {
t.Fatal(err)
}
defer func() { _ = resp2.Body.Close() }()
if resp2.StatusCode != http.StatusOK {
respBody, _ := io.ReadAll(resp2.Body)
t.Fatalf(
"expected 200, got %d: %s",
resp2.StatusCode, respBody,
)
}
}
func TestNickBroadcastToChannels(t *testing.T) {
tserver := newTestServer(t)
aliceToken := tserver.createSession("nick_a")

View File

@@ -2,6 +2,7 @@ package handlers
import (
"encoding/json"
"net"
"net/http"
"strings"
@@ -10,6 +11,33 @@ import (
const minPasswordLength = 8
// clientIP extracts the client IP address from the request.
// It checks X-Forwarded-For and X-Real-IP headers before
// falling back to RemoteAddr.
func clientIP(request *http.Request) string {
if forwarded := request.Header.Get("X-Forwarded-For"); forwarded != "" {
// X-Forwarded-For may contain a comma-separated list;
// the first entry is the original client.
parts := strings.SplitN(forwarded, ",", 2) //nolint:mnd // split into two parts
ip := strings.TrimSpace(parts[0])
if ip != "" {
return ip
}
}
if realIP := request.Header.Get("X-Real-IP"); realIP != "" {
return strings.TrimSpace(realIP)
}
host, _, err := net.SplitHostPort(request.RemoteAddr)
if err != nil {
return request.RemoteAddr
}
return host
}
// HandleRegister creates a new user with a password.
func (hdlr *Handlers) HandleRegister() http.HandlerFunc {
return func(
@@ -82,6 +110,9 @@ func (hdlr *Handlers) handleRegister(
return
}
hdlr.stats.IncrSessions()
hdlr.stats.IncrConnections()
hdlr.deliverMOTD(request, clientID, sessionID, payload.Nick)
hdlr.respondJSON(writer, request, map[string]any{
@@ -134,6 +165,21 @@ func (hdlr *Handlers) handleLogin(
writer http.ResponseWriter,
request *http.Request,
) {
ip := clientIP(request)
if !hdlr.loginLimiter.Allow(ip) {
writer.Header().Set(
"Retry-After", "1",
)
hdlr.respondError(
writer, request,
"too many login attempts, try again later",
http.StatusTooManyRequests,
)
return
}
type loginRequest struct {
Nick string `json:"nick"`
Password string `json:"password"`
@@ -180,6 +226,8 @@ func (hdlr *Handlers) handleLogin(
return
}
hdlr.stats.IncrConnections()
hdlr.deliverMOTD(
request, clientID, sessionID, payload.Nick,
)

View File

@@ -16,6 +16,8 @@ import (
"git.eeqj.de/sneak/neoirc/internal/hashcash"
"git.eeqj.de/sneak/neoirc/internal/healthcheck"
"git.eeqj.de/sneak/neoirc/internal/logger"
"git.eeqj.de/sneak/neoirc/internal/ratelimit"
"git.eeqj.de/sneak/neoirc/internal/stats"
"go.uber.org/fx"
)
@@ -30,6 +32,7 @@ type Params struct {
Config *config.Config
Database *db.Database
Healthcheck *healthcheck.Healthcheck
Stats *stats.Tracker
}
const defaultIdleTimeout = 30 * 24 * time.Hour
@@ -41,6 +44,8 @@ type Handlers struct {
hc *healthcheck.Healthcheck
broker *broker.Broker
hashcashVal *hashcash.Validator
loginLimiter *ratelimit.Limiter
stats *stats.Tracker
cancelCleanup context.CancelFunc
}
@@ -54,12 +59,24 @@ func New(
resource = "neoirc"
}
loginRate := params.Config.LoginRateLimit
if loginRate <= 0 {
loginRate = ratelimit.DefaultRate
}
loginBurst := params.Config.LoginRateBurst
if loginBurst <= 0 {
loginBurst = ratelimit.DefaultBurst
}
hdlr := &Handlers{ //nolint:exhaustruct // cancelCleanup set in startCleanup
params: &params,
log: params.Logger.Get(),
hc: params.Healthcheck,
broker: broker.New(),
hashcashVal: hashcash.NewValidator(resource),
loginLimiter: ratelimit.New(loginRate, loginBurst),
stats: params.Stats,
}
lifecycle.Append(fx.Hook{
@@ -151,6 +168,10 @@ func (hdlr *Handlers) stopCleanup() {
if hdlr.cancelCleanup != nil {
hdlr.cancelCleanup()
}
if hdlr.loginLimiter != nil {
hdlr.loginLimiter.Stop()
}
}
func (hdlr *Handlers) cleanupLoop(ctx context.Context) {

View File

@@ -12,7 +12,7 @@ func (hdlr *Handlers) HandleHealthCheck() http.HandlerFunc {
writer http.ResponseWriter,
request *http.Request,
) {
resp := hdlr.hc.Healthcheck()
resp := hdlr.hc.Healthcheck(request.Context())
hdlr.respondJSON(writer, request, resp, httpStatusOK)
}
}

View File

@@ -10,6 +10,7 @@ import (
"git.eeqj.de/sneak/neoirc/internal/db"
"git.eeqj.de/sneak/neoirc/internal/globals"
"git.eeqj.de/sneak/neoirc/internal/logger"
"git.eeqj.de/sneak/neoirc/internal/stats"
"go.uber.org/fx"
)
@@ -21,6 +22,7 @@ type Params struct {
Config *config.Config
Logger *logger.Logger
Database *db.Database
Stats *stats.Tracker
}
// Healthcheck tracks server uptime and provides health status.
@@ -64,11 +66,22 @@ type Response struct {
Version string `json:"version"`
Appname string `json:"appname"`
Maintenance bool `json:"maintenanceMode"`
// Runtime statistics.
Sessions int64 `json:"sessions"`
Clients int64 `json:"clients"`
QueuedLines int64 `json:"queuedLines"`
Channels int64 `json:"channels"`
ConnectionsSinceBoot int64 `json:"connectionsSinceBoot"`
SessionsSinceBoot int64 `json:"sessionsSinceBoot"`
MessagesSinceBoot int64 `json:"messagesSinceBoot"`
}
// Healthcheck returns the current health status of the server.
func (hcheck *Healthcheck) Healthcheck() *Response {
return &Response{
func (hcheck *Healthcheck) Healthcheck(
ctx context.Context,
) *Response {
resp := &Response{
Status: "ok",
Now: time.Now().UTC().Format(time.RFC3339Nano),
UptimeSeconds: int64(hcheck.uptime().Seconds()),
@@ -76,6 +89,64 @@ func (hcheck *Healthcheck) Healthcheck() *Response {
Appname: hcheck.params.Globals.Appname,
Version: hcheck.params.Globals.Version,
Maintenance: hcheck.params.Config.MaintenanceMode,
Sessions: 0,
Clients: 0,
QueuedLines: 0,
Channels: 0,
ConnectionsSinceBoot: hcheck.params.Stats.ConnectionsSinceBoot(),
SessionsSinceBoot: hcheck.params.Stats.SessionsSinceBoot(),
MessagesSinceBoot: hcheck.params.Stats.MessagesSinceBoot(),
}
hcheck.populateDBStats(ctx, resp)
return resp
}
// populateDBStats fills in database-derived counters.
func (hcheck *Healthcheck) populateDBStats(
ctx context.Context,
resp *Response,
) {
sessions, err := hcheck.params.Database.GetUserCount(ctx)
if err != nil {
hcheck.log.Error(
"healthcheck: session count failed",
"error", err,
)
} else {
resp.Sessions = sessions
}
clients, err := hcheck.params.Database.GetClientCount(ctx)
if err != nil {
hcheck.log.Error(
"healthcheck: client count failed",
"error", err,
)
} else {
resp.Clients = clients
}
queued, err := hcheck.params.Database.GetQueueEntryCount(ctx)
if err != nil {
hcheck.log.Error(
"healthcheck: queue entry count failed",
"error", err,
)
} else {
resp.QueuedLines = queued
}
channels, err := hcheck.params.Database.GetChannelCount(ctx)
if err != nil {
hcheck.log.Error(
"healthcheck: channel count failed",
"error", err,
)
} else {
resp.Channels = channels
}
}

View File

@@ -0,0 +1,122 @@
// Package ratelimit provides per-IP rate limiting for HTTP endpoints.
package ratelimit
import (
"sync"
"time"
"golang.org/x/time/rate"
)
const (
// DefaultRate is the default number of allowed requests per second.
DefaultRate = 1.0
// DefaultBurst is the default maximum burst size.
DefaultBurst = 5
// DefaultSweepInterval controls how often stale entries are pruned.
DefaultSweepInterval = 10 * time.Minute
// DefaultEntryTTL is how long an unused entry lives before eviction.
DefaultEntryTTL = 15 * time.Minute
)
// entry tracks a per-IP rate limiter and when it was last used.
type entry struct {
limiter *rate.Limiter
lastSeen time.Time
}
// Limiter manages per-key rate limiters with automatic cleanup
// of stale entries.
type Limiter struct {
mu sync.Mutex
entries map[string]*entry
rate rate.Limit
burst int
entryTTL time.Duration
stopCh chan struct{}
}
// New creates a new per-key rate Limiter.
// The ratePerSec parameter sets how many requests per second are
// allowed per key. The burst parameter sets the maximum number of
// requests that can be made in a single burst.
func New(ratePerSec float64, burst int) *Limiter {
limiter := &Limiter{
mu: sync.Mutex{},
entries: make(map[string]*entry),
rate: rate.Limit(ratePerSec),
burst: burst,
entryTTL: DefaultEntryTTL,
stopCh: make(chan struct{}),
}
go limiter.sweepLoop()
return limiter
}
// Allow reports whether a request from the given key should be
// allowed. It consumes one token from the key's rate limiter.
func (l *Limiter) Allow(key string) bool {
l.mu.Lock()
ent, exists := l.entries[key]
if !exists {
ent = &entry{
limiter: rate.NewLimiter(l.rate, l.burst),
lastSeen: time.Now(),
}
l.entries[key] = ent
} else {
ent.lastSeen = time.Now()
}
l.mu.Unlock()
return ent.limiter.Allow()
}
// Stop terminates the background sweep goroutine.
func (l *Limiter) Stop() {
close(l.stopCh)
}
// Len returns the number of tracked keys (for testing).
func (l *Limiter) Len() int {
l.mu.Lock()
defer l.mu.Unlock()
return len(l.entries)
}
// sweepLoop periodically removes entries that haven't been seen
// within the TTL.
func (l *Limiter) sweepLoop() {
ticker := time.NewTicker(DefaultSweepInterval)
defer ticker.Stop()
for {
select {
case <-ticker.C:
l.sweep()
case <-l.stopCh:
return
}
}
}
// sweep removes stale entries.
func (l *Limiter) sweep() {
l.mu.Lock()
defer l.mu.Unlock()
cutoff := time.Now().Add(-l.entryTTL)
for key, ent := range l.entries {
if ent.lastSeen.Before(cutoff) {
delete(l.entries, key)
}
}
}

View File

@@ -0,0 +1,106 @@
package ratelimit_test
import (
"testing"
"git.eeqj.de/sneak/neoirc/internal/ratelimit"
)
func TestNewCreatesLimiter(t *testing.T) {
t.Parallel()
limiter := ratelimit.New(1.0, 5)
defer limiter.Stop()
if limiter == nil {
t.Fatal("expected non-nil limiter")
}
}
func TestAllowWithinBurst(t *testing.T) {
t.Parallel()
limiter := ratelimit.New(1.0, 3)
defer limiter.Stop()
for i := range 3 {
if !limiter.Allow("192.168.1.1") {
t.Fatalf(
"request %d should be allowed within burst",
i+1,
)
}
}
}
func TestAllowExceedsBurst(t *testing.T) {
t.Parallel()
// Rate of 0 means no token replenishment, only burst.
limiter := ratelimit.New(0, 3)
defer limiter.Stop()
for range 3 {
limiter.Allow("10.0.0.1")
}
if limiter.Allow("10.0.0.1") {
t.Fatal("fourth request should be denied after burst exhausted")
}
}
func TestAllowSeparateKeys(t *testing.T) {
t.Parallel()
// Rate of 0, burst of 1 — only one request per key.
limiter := ratelimit.New(0, 1)
defer limiter.Stop()
if !limiter.Allow("10.0.0.1") {
t.Fatal("first request for key A should be allowed")
}
if !limiter.Allow("10.0.0.2") {
t.Fatal("first request for key B should be allowed")
}
if limiter.Allow("10.0.0.1") {
t.Fatal("second request for key A should be denied")
}
if limiter.Allow("10.0.0.2") {
t.Fatal("second request for key B should be denied")
}
}
func TestLenTracksKeys(t *testing.T) {
t.Parallel()
limiter := ratelimit.New(1.0, 5)
defer limiter.Stop()
if limiter.Len() != 0 {
t.Fatalf("expected 0 entries, got %d", limiter.Len())
}
limiter.Allow("10.0.0.1")
limiter.Allow("10.0.0.2")
if limiter.Len() != 2 {
t.Fatalf("expected 2 entries, got %d", limiter.Len())
}
// Same key again should not increase count.
limiter.Allow("10.0.0.1")
if limiter.Len() != 2 {
t.Fatalf("expected 2 entries, got %d", limiter.Len())
}
}
func TestStopDoesNotPanic(t *testing.T) {
t.Parallel()
limiter := ratelimit.New(1.0, 5)
limiter.Stop()
}

52
internal/stats/stats.go Normal file
View File

@@ -0,0 +1,52 @@
// Package stats tracks runtime statistics since server boot.
package stats
import (
"sync/atomic"
)
// Tracker holds atomic counters for runtime statistics
// that accumulate since the server started.
type Tracker struct {
connectionsSinceBoot atomic.Int64
sessionsSinceBoot atomic.Int64
messagesSinceBoot atomic.Int64
}
// New creates a new Tracker with all counters at zero.
func New() *Tracker {
return &Tracker{} //nolint:exhaustruct // atomic fields have zero-value defaults
}
// IncrConnections increments the total connection count.
func (t *Tracker) IncrConnections() {
t.connectionsSinceBoot.Add(1)
}
// IncrSessions increments the total session count.
func (t *Tracker) IncrSessions() {
t.sessionsSinceBoot.Add(1)
}
// IncrMessages increments the total PRIVMSG/NOTICE count.
func (t *Tracker) IncrMessages() {
t.messagesSinceBoot.Add(1)
}
// ConnectionsSinceBoot returns the total number of
// client connections since boot.
func (t *Tracker) ConnectionsSinceBoot() int64 {
return t.connectionsSinceBoot.Load()
}
// SessionsSinceBoot returns the total number of sessions
// created since boot.
func (t *Tracker) SessionsSinceBoot() int64 {
return t.sessionsSinceBoot.Load()
}
// MessagesSinceBoot returns the total number of
// PRIVMSG/NOTICE messages sent since boot.
func (t *Tracker) MessagesSinceBoot() int64 {
return t.messagesSinceBoot.Load()
}

View File

@@ -0,0 +1,117 @@
package stats_test
import (
"testing"
"git.eeqj.de/sneak/neoirc/internal/stats"
)
func TestNew(t *testing.T) {
t.Parallel()
tracker := stats.New()
if tracker == nil {
t.Fatal("expected non-nil tracker")
}
if tracker.ConnectionsSinceBoot() != 0 {
t.Errorf(
"expected 0 connections, got %d",
tracker.ConnectionsSinceBoot(),
)
}
if tracker.SessionsSinceBoot() != 0 {
t.Errorf(
"expected 0 sessions, got %d",
tracker.SessionsSinceBoot(),
)
}
if tracker.MessagesSinceBoot() != 0 {
t.Errorf(
"expected 0 messages, got %d",
tracker.MessagesSinceBoot(),
)
}
}
func TestIncrConnections(t *testing.T) {
t.Parallel()
tracker := stats.New()
tracker.IncrConnections()
tracker.IncrConnections()
tracker.IncrConnections()
got := tracker.ConnectionsSinceBoot()
if got != 3 {
t.Errorf(
"expected 3 connections, got %d", got,
)
}
}
func TestIncrSessions(t *testing.T) {
t.Parallel()
tracker := stats.New()
tracker.IncrSessions()
tracker.IncrSessions()
got := tracker.SessionsSinceBoot()
if got != 2 {
t.Errorf(
"expected 2 sessions, got %d", got,
)
}
}
func TestIncrMessages(t *testing.T) {
t.Parallel()
tracker := stats.New()
tracker.IncrMessages()
got := tracker.MessagesSinceBoot()
if got != 1 {
t.Errorf(
"expected 1 message, got %d", got,
)
}
}
func TestCountersAreIndependent(t *testing.T) {
t.Parallel()
tracker := stats.New()
tracker.IncrConnections()
tracker.IncrSessions()
tracker.IncrMessages()
tracker.IncrMessages()
if tracker.ConnectionsSinceBoot() != 1 {
t.Errorf(
"expected 1 connection, got %d",
tracker.ConnectionsSinceBoot(),
)
}
if tracker.SessionsSinceBoot() != 1 {
t.Errorf(
"expected 1 session, got %d",
tracker.SessionsSinceBoot(),
)
}
if tracker.MessagesSinceBoot() != 2 {
t.Errorf(
"expected 2 messages, got %d",
tracker.MessagesSinceBoot(),
)
}
}