Files
silo-server/internal/nodesessions/tracker.go
203a18ae83 feat(observability): OpenTelemetry logs+traces with secret redaction and slog standardization (#290)
* feat(observability): OpenTelemetry logs+traces with secret redaction

Part of #265. Adds opt-in OpenTelemetry (logs + traces) alongside the existing
stderr + opslog pipeline, plus secret redaction on all sinks. Default-off: with
no OTEL_* / SILO_OTEL_ENABLED config, behavior is unchanged.

Bootstrap (internal/telemetry):
- Setup() builds one shared resource, a TracerProvider (parent-based trace-id
  ratio sampler), a LoggerProvider, and the W3C TraceContext+Baggage propagator
  from env. It installs NO MeterProvider — metrics stay on Prometheus, and the
  built-in no-op global MeterProvider keeps the trace instrumentation libs from
  double-emitting. Shutdown is deferred with a flush timeout.
- Logs are bridged via otelslog fan-out (slog.MultiHandler), level-gated by the
  shared LevelVar and best-effort so a failing collector can't break the console
  or DB branches. stderr + opslog stay untouched.

Secret redaction (internal/logredact):
- A slog.Handler masks secret-keyed attributes (password, token, api_key,
  authorization, cookie, ...) — including .With-bound attrs, nested groups,
  secret-keyed group subtrees, and values behind a LogValuer — on the console
  and OTLP sinks, with a no-op fast path when a record has no secret keys.
  opslog.shouldRedact delegates to logredact.SecretKey so all sinks share one
  marker list.

Rotation is infra-managed (no custom file sink): container runtime for stderr,
collector/backend for OTLP, opslog partition-pruning for the DB. Documented in
docs/architecture/observability.md.

Verification: go build ./..., go vet, gofmt -l — clean; go test
./internal/telemetry/ ./internal/logredact/ -race pass.

AI-use disclosure: implemented with AI assistance (Claude Code), including
adversarial reviews that hardened the bootstrap and fixed two redaction leak
paths; reviewed by the author.

* refactor(observability): slog context+component sweep, sloglint gate (phase 3)

Part of #265. Builds on the OTel bootstrap + redaction commit.

Standardizes every log call site onto the context-carrying slog variants so
records correlate with the active OpenTelemetry trace, and locks the standard
in with a machine gate so future code (human- or AI-authored) can't drift back.

- Call-site sweep: converted the remaining slog.<Level>(...) calls to the
  slog.<Level>Context(ctx, ...) form wherever a context.Context is in scope
  (background/init calls with no ctx are left as-is), across 183 files. Applied
  via a type-aware AST codemod. Log levels and message strings are preserved
  verbatim; a component attr (canonical per-package name) is added to direct
  package-level slog calls. Bound-logger calls keep their existing .With
  bindings. The main.go and telemetry package conversions rode with their file
  in the previous commit to keep each file within a single commit.
- Enforcement (.golangci.yml): enable sloglint with context=scope, static-msg,
  key-naming-case=snake, no-mixed-args. After the sweep all four report zero
  violations repo-wide (tests included), so make lint / CI now blocks any
  regression to the non-context form. The gate ships with the sweep because it
  cannot be green until the legacy sites are converted.

Metrics remain on Prometheus; no behavior change to /metrics or Grafana.

Verification: go build ./..., go vet ./..., gofmt -l — clean; sloglint (all 4
rules) 0 violations repo-wide; log levels verified unchanged.

AI-use disclosure: implemented with AI assistance (Claude Code), including the
codemod; reviewed by the author.

* fix(observability): honor per-signal OTLP protocol and secret WithGroup names

Two Codex review findings on PR #290:

- telemetry: OTEL_EXPORTER_OTLP_{TRACES,LOGS}_PROTOCOL now override the
  generic OTEL_EXPORTER_OTLP_PROTOCOL per signal, so mixed collector
  setups (e.g. HTTP logs + gRPC traces) build the right exporter.
- logredact: entering a group whose name is secret-bearing (e.g.
  WithGroup("authorization")) now masks every leaf in that subtree,
  matching how slog.Group("authorization", ...) is masked as a whole.

* fix(observability): address review feedback on telemetry bootstrap

- Telemetry setup failure no longer kills boot: Setup returns usable
  no-op providers alongside the error and main logs and continues with
  telemetry disabled, honoring the best-effort contract.
- Honor OTEL_TRACES_SAMPLER (always_on/off, traceidratio, parentbased_*
  variants); unsupported values fall back to parentbased_traceidratio.
- Attach node identity as semconv service.instance.id instead of the
  non-semconv node.name.
- Rename opslog retention-scope log attrs to target_component/target_level
  so they no longer collide with the canonical component routing key, and
  tag those lines with component=opslog.
- Fix stale levelGated comment casing; use WarnContext in the telemetry
  shutdown defer; document the LogValuer double-resolve on the redaction
  slow path.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Quick <31828688+Quick104@users.noreply.github.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-07-09 08:53:52 -04:00

258 lines
7.0 KiB
Go

package nodesessions
import (
"context"
"crypto/sha256"
"encoding/hex"
"encoding/json"
"log/slog"
"sync"
"time"
"github.com/redis/go-redis/v9"
)
const (
keyPrefix = "silo:sessions:"
sessionTTL = 60 * time.Second
refreshInt = 30 * time.Second
)
// SessionInfo represents an active streaming session stored in Redis.
type SessionInfo struct {
SessionID string `json:"session_id"`
NodeURL string `json:"node_url"`
NodeName string `json:"node_name"`
UserID string `json:"user_id,omitempty"`
MediaItemID string `json:"media_item_id,omitempty"`
MediaTitle string `json:"media_title,omitempty"`
Type string `json:"type"` // "direct_play", "remux", "transcode"
CodecVideo string `json:"codec_video,omitempty"`
CodecAudio string `json:"codec_audio,omitempty"`
Resolution string `json:"resolution,omitempty"`
HWAccel string `json:"hw_accel,omitempty"`
StartedAt string `json:"started_at"`
// AuthUserID / ProfileID / MediaFileID are the numeric ownership keys the
// node copies from the verified stream token. They enrich the live admin
// "active streams" view (served by SCANning these records) so it can answer
// *who* is watching *what* on each node, not just session id + node + type;
// the string UserID/MediaItemID/MediaTitle fields remain the display labels.
AuthUserID int `json:"auth_user_id,omitempty"`
ProfileID string `json:"profile_id,omitempty"`
MediaFileID int `json:"media_file_id,omitempty"`
}
// Tracker manages session lifecycle in Redis for a single node.
type Tracker struct {
rdb *redis.Client
nodeURL string
nodeName string
nodeType string
nodeHash string // first 8 chars of SHA-256 of nodeURL
mu sync.Mutex
sessions map[string]struct{} // set of active session IDs
touched map[string]time.Time // ephemeral sessions by last-activity time
}
// NewTracker creates a session tracker for the given node.
// rdb may be nil, in which case all operations are no-ops.
func NewTracker(rdb *redis.Client, nodeURL, nodeName, nodeType string) *Tracker {
h := sha256.Sum256([]byte(nodeURL))
return &Tracker{
rdb: rdb,
nodeURL: nodeURL,
nodeName: nodeName,
nodeType: nodeType,
nodeHash: hex.EncodeToString(h[:4]), // 8 hex chars
sessions: make(map[string]struct{}),
touched: make(map[string]time.Time),
}
}
// redisKey returns the full Redis key for a session.
func (tr *Tracker) redisKey(sessionID string) string {
return keyPrefix + tr.nodeHash + ":" + sessionID
}
// NodeHash returns the node's hash prefix used in Redis keys.
func (tr *Tracker) NodeHash() string {
return tr.nodeHash
}
// NodeURL returns the node's URL.
func (tr *Tracker) NodeURL() string {
return tr.nodeURL
}
// NodeName returns the node's display name.
func (tr *Tracker) NodeName() string {
return tr.nodeName
}
// ActiveCount returns the number of active sessions tracked by this node,
// including ephemeral sessions touched within the session TTL.
func (tr *Tracker) ActiveCount() int {
tr.mu.Lock()
defer tr.mu.Unlock()
now := time.Now()
count := len(tr.sessions)
for id, last := range tr.touched {
if _, dup := tr.sessions[id]; dup {
continue
}
if now.Sub(last) <= sessionTTL {
count++
}
}
return count
}
// Track registers an active session in Redis with a TTL.
func (tr *Tracker) Track(ctx context.Context, info SessionInfo) {
if tr.rdb == nil {
return
}
data, err := json.Marshal(info)
if err != nil {
slog.DebugContext(ctx, "session track marshal failed", "component", "nodesessions", "error", err)
return
}
key := tr.redisKey(info.SessionID)
if err := tr.rdb.Set(ctx, key, data, sessionTTL).Err(); err != nil {
slog.DebugContext(ctx, "session track set failed", "component", "nodesessions", "error", err, "session", info.SessionID)
return
}
tr.mu.Lock()
tr.sessions[info.SessionID] = struct{}{}
tr.mu.Unlock()
}
// Touch registers or refreshes an ephemeral session that has no explicit end,
// such as HLS manifest/segment fetches flowing through a proxy. The session is
// written to Redis on first touch and drops out of the active count after
// sessionTTL without further touches (pruned by the refresh loop).
func (tr *Tracker) Touch(ctx context.Context, info SessionInfo) {
if tr.rdb == nil {
return
}
tr.mu.Lock()
_, known := tr.touched[info.SessionID]
tr.touched[info.SessionID] = time.Now()
tr.mu.Unlock()
if known {
return
}
data, err := json.Marshal(info)
if err != nil {
slog.DebugContext(ctx, "session touch marshal failed", "component", "nodesessions", "error", err)
return
}
if err := tr.rdb.Set(ctx, tr.redisKey(info.SessionID), data, sessionTTL).Err(); err != nil {
slog.DebugContext(ctx, "session touch set failed", "component", "nodesessions", "error", err, "session", info.SessionID)
}
}
// Remove deletes a session from Redis and the in-memory set.
func (tr *Tracker) Remove(ctx context.Context, sessionID string) {
if tr.rdb == nil {
return
}
tr.mu.Lock()
delete(tr.sessions, sessionID)
delete(tr.touched, sessionID)
tr.mu.Unlock()
if err := tr.rdb.Del(ctx, tr.redisKey(sessionID)).Err(); err != nil {
slog.DebugContext(ctx, "session remove failed", "component", "nodesessions", "error", err, "session", sessionID)
}
}
// Cleanup deletes all session keys for this node. Called on graceful shutdown.
func (tr *Tracker) Cleanup(ctx context.Context) {
if tr.rdb == nil {
return
}
tr.mu.Lock()
ids := make([]string, 0, len(tr.sessions)+len(tr.touched))
for id := range tr.sessions {
ids = append(ids, id)
}
for id := range tr.touched {
if _, dup := tr.sessions[id]; !dup {
ids = append(ids, id)
}
}
tr.sessions = make(map[string]struct{})
tr.touched = make(map[string]time.Time)
tr.mu.Unlock()
if len(ids) == 0 {
return
}
pipe := tr.rdb.Pipeline()
for _, id := range ids {
pipe.Del(ctx, tr.redisKey(id))
}
if _, err := pipe.Exec(ctx); err != nil {
slog.DebugContext(ctx, "session cleanup pipeline failed", "component", "nodesessions", "error", err)
}
}
// StartRefresh starts a background goroutine that refreshes TTLs for all
// active sessions every 30 seconds. Stops when ctx is cancelled.
func (tr *Tracker) StartRefresh(ctx context.Context) {
if tr.rdb == nil {
return
}
go func() {
ticker := time.NewTicker(refreshInt)
defer ticker.Stop()
for {
select {
case <-ctx.Done():
return
case <-ticker.C:
tr.refreshAll(ctx)
}
}
}()
}
func (tr *Tracker) refreshAll(ctx context.Context) {
now := time.Now()
tr.mu.Lock()
ids := make([]string, 0, len(tr.sessions)+len(tr.touched))
for id := range tr.sessions {
ids = append(ids, id)
}
for id, last := range tr.touched {
if now.Sub(last) > sessionTTL {
// Idle ephemeral session: stop refreshing and let the Redis
// key expire on its own.
delete(tr.touched, id)
continue
}
if _, dup := tr.sessions[id]; !dup {
ids = append(ids, id)
}
}
tr.mu.Unlock()
if len(ids) == 0 {
return
}
pipe := tr.rdb.Pipeline()
for _, id := range ids {
pipe.Expire(ctx, tr.redisKey(id), sessionTTL)
}
if _, err := pipe.Exec(ctx); err != nil {
slog.DebugContext(ctx, "session refresh pipeline failed", "component", "nodesessions", "error", err)
}
}