Once the database passed about 4.5 GiB every stats request ran a COUNT(*)
over each table plus a MIN/MAX union scan of both route tables, took the
full timeout and returned HTTP 500, so the status page went blank.
The server now keeps the last database statistics in memory and recomputes
them at most once every 30 seconds; requests serve the cached copy and a
stale copy triggers a single background refresh, so no request runs the
scans. The route-count split is folded into the cached stats, removing the
separate per-request live-route count query.
The oldest/newest route timestamps now read one row from each end of the
last_updated index instead of scanning both tables, and select the column
directly so the driver parses it into time.Time; the old aggregate returned
an untyped string that failed to scan and logged a warning every call.
Model: opus-4-8
Batch flush paths read before they write, so a deferred transaction starts
as a reader and must upgrade to the write lock on its first INSERT/UPDATE.
When the background maintainer holds the write lock for a WAL checkpoint,
that upgrade fails immediately with "database is locked" and the busy
timeout does not apply, so the batch is dropped. Adding _txlock=immediate
to the DSN makes every transaction take the write lock at BEGIN, so it
waits up to busy_timeout instead of failing.
A regression test drives batch writes against a running checkpoint loop and
fails with "database is locked" without the change.
Model: opus-4-8
After the earlier memory work, SQLite's live heap is bounded but process
RSS still climbed about 70 MiB/min, nearly all anonymous and outside the
Go runtime and outside SQLite's own accounting, without any SQLITE_NOMEM.
The cause is glibc: the SQLite C library allocates and frees millions of
small page-cache chunks from many threads, and glibc keeps each arena's
freed chunks resident. With arenas uncapped it creates up to eight per
core, so on a large host RSS grows with the core count.
Set MALLOC_ARENA_MAX=2 in the image to bound the retained memory. Writes
are already serialized, so the two-arena cap costs no throughput. Update
the README Memory section.
Model: opus-4-8
Add ENV GOMEMLIMIT=1536MiB to the runtime stage so the Go runtime keeps a
1.5 GiB soft heap limit. runuser preserves it the way it already does
XDG_DATA_HOME, so the routewatch process inherits it.
Add a Memory section to the README describing the budget (Go 1.5 GiB soft,
SQLite 640 MiB pool cache and 1.5 GiB hard heap, ~0.2 GiB other), the
required 5 GiB container limit, how to override GOMEMLIMIT, what happens at
each limit, and the DEBUG=routewatch System stats line. Every sentence
matches the behaviour already merged to next.
Model: opus-4-8
Each handler queue held 100,000 message pointers; a message is retained
until the slowest handler drains it, so all four full was a derived worst
case near 800 MiB. Twenty thousand is about four seconds of feed at peak
and caps that at roughly 160 MiB. The streamer already drops rather than
blocks on a full queue, so the smaller bound is safe.
Batch sizes are unchanged. The largest, asnBatchSize, is 30,000 and now
exceeds its queue, but each queued message contributes every ASN in its
path, and every handler also flushes on its own timer regardless of fill,
so batches still flush and no size change is warranted.
Model: opus-4-8
The stats handlers ran the database query in a goroutine that sent on an
unbuffered channel. When the 4s request timeout won, nothing received and
the goroutine blocked forever; the status page polls every 2s, so once the
query exceeds the timeout every poll leaked one goroutine. Give both
channels capacity 1 so the send always completes.
The streamer started two ticker goroutines per connection that exited only
with the streamer's lifetime context, leaking two on every reconnect. Scope
them to a per-connection context cancelled when the stream call returns.
Tests force the stats timeout repeatedly and drive many reconnects, then
assert the goroutine count settles back to its starting value. The streamer
gains an internal endpoint field so a test can point it at a local server.
Model: opus-4-8
Community and Raw are decoded from every live message and read by no
handler; Raw is the hex of the whole BGP message. Each parsed message
sits in up to four handler queues, so retaining them is a large share of
queue memory. Tagging both json:"-" keeps them out of the decoded
message while the fields handlers use still decode. A decode test over
docs/message-examples.json confirms both stay empty and the used fields
(Path, Announcements) still populate.
Model: opus-4-8
Copy .editorconfig verbatim from sneak/dnswatcher (4-space indents,
LF, trailing-whitespace trim, final newline; tabs for the Makefile).
Add .gitea/workflows/check.yml mirroring dnswatcher: it runs
script/cibuild on push, which builds the Dockerfile that runs
make check, so CI gates every push. The checkout action is pinned by
commit SHA as that repo pins it. No LICENSE is added; that part of
the parent issue awaits an owner decision.
Model: opus-4-8
processPeerings now takes the accumulated AS-path map under the lock and
replaces it with a fresh empty one, so each path is processed once and the
map never grows past a single 30-second interval's traffic. HandleMessage
stops adding new paths once maxTrackedPaths (500000) is reached and counts
the drops, which are logged with each run. The 30-minute prune and its
ticker are removed as dead code under the swap, and the map mutex drops
from RWMutex to Mutex since the read path is gone. RecordPeering already
upserts last_seen, so stored peerings are unchanged.
Model: opus-4-8
The 3 GiB page cache and temp_store=MEMORY were set once in Initialize, so
only one pooled connection carried them and the other nine got no busy_timeout,
which drove many "database is locked" errors.
Move the per-connection settings into the DSN so every pooled connection gets a
64 MiB cache (640 MiB worst case over ten connections), synchronous OFF, a
5 s busy_timeout and WAL. Drop those pragmas from Initialize; DISTINCT temp
B-trees now spill to disk. Add process-wide soft (1 GiB) and hard (1.5 GiB)
heap limits; at the hard limit a statement returns SQLITE_NOMEM and the
existing batch paths log and drop, so nothing panics or exits.
Model: opus-4-8
REPO_POLICIES.md requires the container build to run the checks; the
Dockerfile only compiled the binary, so script/cibuild checked nothing. A
new lint stage on the golangci-lint v2.7.2 image runs make fmt-check and
make lint, and the build stage waits for it and then runs make test, so
make docker now fails when formatting, lint or a test fails. All three base
images are pinned by digest with their versions unchanged. The final image
and how it starts are unchanged.
make docker is now the gate to use on hosts where the installed linter is
older than the Go toolchain and make lint cannot run directly.
Model: opus-4-8 (implementation, review); fable-5-1 (summary)
TestRouteWatchLiveFeed streams the live RIPE RIS feed for a few seconds, so
the default test run depended on the network and tripped the race detector.
It now skips under testing.Short(), and script/test passes -short in both
go test lines, so make check runs offline and gives the same result every
time. The README says how to run the live test on demand. No non-test code
changed.
The race itself is not fixed: the test reads the peeringHandler field of
RouteWatch while Run is still setting it. Nothing outside the test reads
that field concurrently.
Model: opus-4-8 (implementation, review); fable-5-1 (summary)
PASSIVE checkpoints may not fully checkpoint when writers are active.
TRUNCATE mode blocks writers briefly but ensures complete checkpointing,
keeping the WAL small for consistent read performance under heavy load.
Under heavy write load, 30 seconds is too long between checkpoints,
causing the WAL to grow and slow down read queries. More aggressive
checkpointing keeps the WAL small and maintains read performance.
Explicitly checkpoint the WAL on database initialization to consolidate
any large WAL file from previous runs. Log the checkpoint results to
help diagnose issues.
- Log slow requests (>1s) at WARNING level with slow=true flag
- Log request timeouts at WARNING level in TimeoutMiddleware
- Replace heavy GetStatsContext with lightweight Ping in health check
- Add Ping method to database interface (SELECT 1)
The WAL file was growing to 700MB+ which caused COUNT(*) queries to
timeout. Reads must scan the WAL to find current page versions, and
a large WAL makes this slow.
Add Checkpoint method to database interface and run PASSIVE checkpoints
every 30 seconds via the DBMaintainer. This keeps the WAL small and
maintains fast read performance under heavy write load.
Use runuser to drop privileges and execute the app as the routewatch
user (uid 1000). Fix data directory permissions at runtime since host
mounts may have incorrect ownership.
Display oldest and newest route timestamps in the Routing Table card
on the /status page. Timestamps are shown as relative times (e.g.,
"5m ago", "2h 30m ago").
Changes:
- Add OldestRoute and NewestRoute fields to database.Stats
- Query MIN/MAX of last_updated across live_routes_v4 and v6 tables
- Include timestamps in both /status.json and /api/v1/stats responses
- Add formatRelativeTime JavaScript function for display
- Create new home page (/) with overview stats, ASN lookup,
AS name search, and IP address lookup with JSON display
- Add responsive navbar to all pages with app branding
- Navbar shows "routewatch by @sneak" with link to author
- Status page accessible via navbar link
- Remove redirect from / to /status, serve home page instead
Change route from wildcard /prefix/* to explicit /prefix/{prefix}/{len}
to properly handle URL-encoded IPv6 addresses with CIDR notation.
- Separate prefix and length into individual path parameters
- Add prefixURL template function for generating correct links
- Remove url.QueryUnescape from handlers (chi handles decoding)
- Add GC statistics (run count, total/last pause, heap usage)
- Add BGP peer count tracking from RIS Live OPEN/NOTIFICATION messages
- Add route churn rate metric (announcements + withdrawals per second)
- Add announcement and withdrawal counters
- Add footer with attribution, license, and git revision
- Embed git revision at build time via ldflags
- Update HTML template to display all new metrics
- Track reconnection count in metrics tracker
- Display connection duration under Stream Statistics
- Display reconnect count since app startup
- Update both JSON API and HTML status page
- Replace chi's Logger middleware with structured slog-based logging
- Log request start (debug) and completion (info/warn/error by status)
- Include method, path, status, duration_ms, remote_addr in logs
- Increase request timeout from 8s to 30s for slow queries
- Add read/write/idle timeouts to HTTP server config
- Better server startup logging to confirm listening state
- Builder stage: vendor dependencies, build binary, create source archive
- Source archive (.tar.zst) includes all code and vendored dependencies
- Runtime stage: minimal Debian image with binary and source archive
- Health check via curl to /.well-known/healthcheck.json
- Runs as non-root user (routewatch:1000)
- Use PRAGMA incremental_vacuum instead of full VACUUM
- Frees ~1000 pages (~4MB) per run without blocking writes
- Run every 10 minutes instead of 6 hours since it's lightweight
- Set auto_vacuum=INCREMENTAL pragma for new databases
- Remove blocking VACUUM on startup
- Add health check endpoint at /.well-known/healthcheck.json that
verifies database and RIS Live connectivity, returns 200/503
- Fix panic in streamer when encountering unknown RIS message types
by logging a warning and continuing instead of crashing
- Add DBMaintainer for periodic database maintenance:
- VACUUM every 6 hours to reclaim space
- ANALYZE every hour to update query statistics
- Graceful shutdown support
- Add Vacuum() and Analyze() methods to database interface
- Add WHOIS Fetcher card showing fresh/stale/never-fetched ASN counts
- Display hourly success/error counts and current fetch interval
- Increase max WHOIS rate to 1/sec (down from 10 sec minimum)
- Select random stale ASN instead of oldest for better distribution
- Add index on whois_updated_at for query performance
- Track success/error timestamps for hourly stats
- Add GetWHOISStats database method for freshness statistics
- Always return consistent JSON structure with query and results array
- Add PTR field to IPInfo for reverse DNS records
- Support comma-separated IPs and hostnames in single query
- Do PTR lookup for all IPs (direct, resolved from hostname, or listed)
- Remove trailing dots from PTR records
- Accept hostnames in addition to IP addresses for /ip endpoints
- Resolve A and AAAA records for hostnames
- Return list of results with info for each resolved IP
- Include hostname in response when resolving hostnames
- Report per-IP errors while still returning successful lookups
- Reduce base interval from 60s to 15s for faster initial fetching
- Add exponential backoff on failure (up to 5 minute max interval)
- Decrease interval on success (down to 10 second minimum)
- Add mutex to prevent concurrent WHOIS fetches
- Track consecutive failures for backoff calculation
- Add /ip and /ip/{addr} JSON endpoints returning comprehensive IP info
- Include ASN, netblock, country code, org name, abuse contact, RIR data
- Extend ASN schema with WHOIS fields (country, org, abuse contact, etc)
- Create background WHOIS fetcher for rate-limited ASN info updates
- Store raw WHOIS responses for debugging and data preservation
- Queue on-demand WHOIS lookups when stale data is requested
- Refactor handleIPInfo to serve all IP endpoints consistently
The stream stats were showing decompressed data sizes, not actual wire
bandwidth. This change adds wire byte tracking by disabling automatic
gzip decompression in the HTTP client and wrapping the response body
with a counting reader before decompression. Both wire (compressed) and
decompressed bytes are now tracked and exposed in the API responses.
Add comprehensive godoc comments to all exported types, functions,
and constants throughout the codebase. Create README.md documenting
the project architecture, execution flow, database schema, and
component relationships.
Enhance documentation comments for constants, types, and exported methods
in peeringhandler.go to follow Go documentation conventions. The improved
comments provide more context about the purpose and behavior of each item.
Expand the documentation comment for CLIEntry to provide more context
about what the function does, including its use of the fx dependency
injection framework, signal handling, and blocking behavior.
sneak
clawbot