Implement Request-ID middleware that generates or propagates X-Request-ID
headers for distributed request tracing.
- Check incoming X-Request-ID header, reuse if present (trust downstream)
- Generate UUID v4 via google/uuid if no incoming ID
- Store ID in ctx.UserValue for variable system and access log access
- Set X-Request-ID response header for client-side tracing
- Add GetRequestID() helper for proxy header propagation
- Registered as first middleware (before AccessLog) so $request_id
is available throughout the request lifecycle
- 8 unit tests covering generation, propagation, empty header, UUID format
- handler/static.go: add sync.RWMutex to StaticHandler; protect Handle
with RLock and all setters with Lock to prevent data races
- http2/server.go: delete empty connection slice keys from pool map to
prevent memory leak under high client churn
- loadbalance/slow_start.go: recreate stopCh in Start() to support
Start-Stop-Start cycles; guard double-close in Stop()
- resolver/resolver.go: recreate stopCh in Start() to support restart
- logging/logging.go: save *os.File handles from getOutput so Close()
actually closes log files; exclude os.Stdout/os.Stderr from closing
- ssl/session_tickets.go: protect started/rotateTimer access in
scheduleRotation with mu; support Start-Stop-Start cycles
- ssl/ssl.go: cache parsed default certificate to avoid re-parsing on
every TLS handshake for OCSP stapling
- proxy/proxy.go: decrement connection count on dangerous path rejection
(line 724) to prevent connection count leak
- handler/sendfile_linux.go: return *os.File from getSocketFile and let
linuxSendfile close it, fixing EBADF from deferred close in getSocketFd
- proxy/websocket.go: return bufio.Reader from readWebSocketUpgradeResponse
and wrap targetConn with bufferedConn to consume pre-buffered frame data,
preventing first-frame loss
- server/pool.go: use non-blocking send after starting new worker to avoid
deadlock when queue is full
- stream/stream.go: check stopCh on non-timeout UDP read errors to prevent
infinite loop and shutdown deadlock
- middleware/ratelimit: replace select-based close guard with sync.Once in
StopCleanup to prevent double-close panic
- 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
- logging: pre-allocate fresh slice for request field to avoid mutating
fasthttp internal buffers via append into slices with excess capacity
- mimeutil: move defaultMIME fallback before cache insertion so unknown
extensions are consistently cached as application/octet-stream
- builtin: use -0700 timezone format instead of hardcoded +0800; cache
generated request_id in UserValue to prevent different IDs per expansion
- variable: initialize maps in fallback Context to prevent nil map panic
Replace concurrent polling of srv.GetListeners() with pre-created
net.Listener injected via srv.SetListeners(). The previous approach
raced with the Start() goroutine writing to s.listeners and calling
net.Listen().
Also remove unused waitForServerRunning call.
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: /.
The //go:generate go test -v ./... directive is not what
go:generate is intended for (code generation, not test execution).
Remove it to avoid confusion.
Add netutil.FormatRemoteAddr() to avoid repeated net.TCPAddr.String()
allocations on the hot logging path.
- IPv4 addresses use a zero-allocation fast path (custom uint8->ASCII).
- IPv6 falls back to addr.String() with a 1024-entry LRU cache.
- Update logging.LogAccess and variable $remote_addr/$remote_port
getters to use the shared helper.
Eliminates net.JoinHostPort and net.IP.String from top allocators
when access logging is active.
- Router now enables FileInfoCache with 2s TTL for all static handlers
to reduce redundant os.Stat calls.
- FileInfoCache supports negative caching: missing files are cached
with a shorter TTL to avoid repeated stat on non-existent paths.
- Fix missing fileCache lookup for directory index files (index.html).
Previously handleStandard/handleTryFiles skipped fileCache when
serving index files, causing os.ReadFile on every request even
with file_cache configured.
- Extract tryServeFromFileCache() helper to unify cache hit logic
across file and index-file serving paths.
Verified with wrk 200 conn / 20s on static /index.html:
- Throughput: 140k -> 242k req/sec (+73%)
- alloc_space: 2.6 GB -> 4.6 MB (-99.8%)
Add logging.access.sample_rate config (0.0-1.0) for deterministic
request sampling. 5xx errors are always logged; 2xx/3xx/4xx follow
the configured rate. Uses atomic.Uint64 counter for lock-free,
zero-allocation sampling decisions.
Includes test updates to verify:
- sample_rate=1.0 logs all requests
- sample_rate=0.0 logs only 5xx
- 5xx are always logged regardless of rate
Production static file serving now uses FileInfoCache by default
with a 2-second TTL in router.go, dramatically reducing os.Stat
syscalls for missing files and repeated paths.
Changes:
- Add negative cache support to FileInfoCache (caches 'not found' results)
- Introduce statWithCache() helper in StaticHandler for uniform caching
- Make FileInfoCache TTL configurable via SetTTL()
- Default cacheTTL=0 disables caching in NewStaticHandler (tests compat)
- router.go enables fileInfoCache with 2s TTL for all static handlers
Benchmark (repeated 404s):
No cache: ~2651 ns/op, 2225 B/op, 15 allocs/op
With cache: ~1505 ns/op, 1905 B/op, 12 allocs/op
Improvement: -43% latency, -14% allocations
This addresses the dominant allocation source in v0.4.0 profile
(os.statNolog at 74.95% of allocations).
- Add writeAllocsProfile() helper in pprof_impl.go
- Register /allocs route in PprofHandler.ServeHTTP
- Add handleAllocs() method with proper streaming response
- Update index page to list the new allocs profile link
This aligns lolly's pprof endpoints with net/http/pprof and enables
allocation hotspot analysis during performance benchmarking.
- Collect baseline benchmark summary across all core modules
- Save key results to benchmarks/v0.4.0/summary.txt
- Update .gitignore to track benchmark summaries/reports
- Include performance optimization design docs and plan
- Fix handleConnection to use addr parameter for direct upstream map
lookup instead of always selecting the first upstream
- Add Server.Stop() for graceful shutdown with listener closing, UDP
server cleanup, health checker termination, and goroutine joining
- Add shutdownStream() to App and call it in SIGTERM/SIGQUIT/SIGUSR2
signal handlers to prevent goroutine and port leaks on shutdown
- Verify Least Time picks faster target consistently
- Verify Sticky fallback when target becomes unhealthy
- Test cookie encoding and session persistence
- Record headerTime when header is received
- Record lastByteTime when response is complete
- Use correct timing calculations (headerReceived/connectEnd/responseEnd)
- Add least_time and sticky to createBalancerByName
- Implement response time recording for Least Time
- Support StickySession in target selector with request context
- StickySession auto-starts when created
- Add least_time and sticky to valid algorithms list
- Add LeastTimeConfig and StickyConfig structures
- Update default config generation with new options
- Add configuration validation for new fields
- Add sync.Once to prevent double close of stopCh in Stop()
- Add nil fallback guard in NewStickySession (defaults to RoundRobin)
- Add atomic.Bool to make Start() idempotent
- Add tests for double Stop() and nil fallback scenarios
- Fix Select to check if cookie is expired before routing
- Add TestStickySession_ExpiredCookie test
- Expired cookies now trigger fallback + new cookie set
- Encode cookie as base64(target_url + | + timestamp) per spec
- Use cookie value (not targetURL) for shard key and session map keys
- Add missing sticky.Start() calls in tests
- Fix time precision in cookie encode/decode tests
- Add atomic EWMA Stats field to Target
- Implement LeastTime balancer with header_time and last_byte metrics
- Support Select and SelectExcluding with zero-lock design
- Add ResponseTimeRecorder interface for proxy integration
- Zero-lock atomic EWMA implementation using fixed-point arithmetic
- Supports header_time and last_byte_time tracking
- Concurrent-safe with CAS retry loop
