Batch 1 - Crash/data corruption:
- Add nil/empty config guards in server and app
- Protect against fasthttp buffer mutation in logging
- Ensure mimeutil falls back before caching empty MIME
- Make proxy health checker and resolver tolerate nil cfg
- Prevent Lua EMERG/ALERT/CRIT log levels from killing process
Batch 2 - Concurrency/race conditions:
- Add focused race tests for compression pool, ssl/ocsp,
server proxies slice, Lua timer cancel, and TCP socket connect
- Fix Lua timer active race and socket ConnectAsync race
Batch 3 - Resource leaks/functional damage:
- Track and stop rate limiter cleanup goroutines on shutdown
- Add tests for proxy connection count leaks, server pool deadlock,
Linux sendfile integrity, WebSocket frame data loss,
stream UDP stop deadlock, and upstream name mismatch
- Fix stream ListenTCP/upstream lookup and UDP shutdown
Batch 4 - High severity technical debt:
- Add nil guard in security headers middleware
- Validate sliding window divisor and add tests
- Protect static handler fields with existing RWMutex
- Avoid mutating live HostClient.Addr on DNS updates
- Fix slow start Start/Stop and resolver restart
- Initialize variable fallback context maps
- Add tests for request_id/time_local and logging file handle close
Also:
- Relax integration variable performance threshold under race detector
- Update benchmark plan docs for new stream ListenTCP signature
Verification:
- make test: pass
- go test -race -count=1 ./internal/...: pass
- make lint: 0 issues
- make build: success
Add lightweight health check endpoints for container orchestration
(Kubernetes liveness/readiness probes, load balancer health checks).
New config under monitoring:
- healthz.enabled (default true), healthz.path (default /healthz)
- readyz.enabled (default true), readyz.path (default /readyz)
/healthz (liveness):
- Always returns 200 {"status":"ok"} if process is alive
- No dependency checks, minimal overhead
/readyz (readiness):
- Returns 200 {"status":"ready"} when server is running
- Returns 503 {"status":"not ready","reasons":[...]} when not ready
- Static-only servers (no proxies) always return 200
Registration:
- Registered alongside status/pprof endpoints
- Available in single mode (LocationEngine) and multi-server mode (Router)
- No IP allowlist required (K8s probes come from localhost)
- 6 unit tests covering all response scenarios
- compression: move sync.Pool.New initialization into constructors to
eliminate lazy-init race in Get()
- ssl/ocsp: copy response fields under RLock before releasing, preventing
race with concurrent writers in refreshAll
- server: change proxiesMu from sync.Mutex to sync.RWMutex; protect
getProxyCacheStats and purge handlers with RLock to prevent races
with proxy registration
- lua/api_timer: fix double-decrement race in Cancel vs executeTimer
by using timer.Stop() result to determine who decrements active
- lua/api_socket_tcp: fix nil pointer race in ConnectAsync by checking
currentOp under lock before Connect returns
- server: reject Start() when config is nil to prevent panic
- app_common: guard empty Servers slice in initHTTP2/3 and logServerAddresses
- proxy/health: handle nil HealthCheckConfig with defaults
- resolver: handle nil ResolverConfig by returning noopResolver
- middleware/headers: skip UpdateConfig when cfg is nil
- middleware/sliding_window: enforce minimum window duration of 1s
- lua/api_log: map EMERG/ALERT/CRIT to Error() instead of Fatal()
to prevent Lua scripts from killing the entire server process
Two related fixes that must land together:
1. config.Load() now starts from DefaultConfig() before unmarshaling
YAML. This ensures missing top-level fields (Performance,
Monitoring, Resolver) use their documented defaults instead of
zero values. Most importantly, file_cache is no longer silently
disabled when users omit the performance: section.
2. startSingleMode() now checks Monitoring.Status.Enabled instead of
Path/Allow to decide whether to register the status endpoint.
Without this change, fix#1 would have caused a regression where
the status handler is registered even when monitoring is disabled,
because DefaultConfig() sets Path and Allow defaults.
Also replace remaining log.Printf in status.go and lua/api_timer.go
with zerolog to follow project logging conventions.
Added tests:
- config/load_test.go: verifies defaults are applied, explicit values
override defaults, and monitoring stays disabled by default.
- server/monitoring_registration_test.go: verifies /_status is only
registered when enabled and remains reachable with static handler
on path: /.
Add matcher.ReleaseMatchResult(result) in the base handler to prevent
sync.Pool object leak. Every Match() call acquires from pool but the
caller never returned objects, causing unbounded pool growth.
- Fix FD leak in DupListener: close *os.File after net.FileListener
- Add cleanup of partially-duped listeners on DupListener failure
- Make reload timeout configurable via shutdown.reload_timeout
- Handle filepath.Abs errors in processIncludes instead of ignoring
- Use net.ParseIP in isAnyAddr for robust IPv6 support
Remove VHost fallback during graceful upgrade. Serialize listener
creation before parallel router/middleware setup to prevent concurrent
inherited listener consumption. Fix tcpAddrMatch to match when either
side is any-addr (0.0.0.0/::).
createListener now checks pre-set s.listeners (Path 2) for hot reload,
not just upgradeManager.IsChild() (Path 1). Add DupListener to dup FDs
so old/new servers own independent listeners. Reload rebuilds HTTP/2
and HTTP/3. Add matchInheritedListener with TCP any-addr matching.
Add requiresFullRestart with VHost server count detection.
Add typed ConflictError for path conflicts, change register functions
to return errors, handle conflicts as warnings and fatal errors as
startup failures. Remove all 20 instances of ignored Add* return values.
Server.running was a plain bool accessed from multiple goroutines
(start/stop/signal handlers). Convert to atomic.Bool with
Store/Load to make all accesses safe for concurrent use.
Updates all test files to use the new atomic API.
- Add nolint comments for type assertion errcheck in gjson/encode.go
(switch case guarantees type safety)
- Handle fasthttp.Serve errors in benchmark mock backends
- Rename error variables to avoid shadowing in server.go
- Use underscore for unused loop variables
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Add registerLuaRoutes method for router-based route registration
and call it in startMultiServerMode to fix Lua routes not working
when multiple servers are configured.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Add call to registerLuaRoutesWithLocationEngine between proxy and static
route registration, ensuring correct routing order: proxy -> lua -> static.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
In startMultiServerMode, status and pprof handlers were not registered,
causing /_status and /debug/pprof endpoints to return 404. Now these
handlers are registered on the server with default: true, consistent
with startVHostMode behavior. Also fixed cache API registration to
use default server instead of first server.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Conflict: sendfile.go (!linux build tag) was incorrectly modified to
include linuxSendfile and getSocketFd functions which already exist
in sendfile_linux.go.
Resolution: Keep HEAD version (simple fallback returning ENOTSUP) as
Linux implementation is properly separated in sendfile_linux.go.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
