fix(server): plug additional listener leaks identified during #782 hunt

[buremba]

Summary

Refs #782.

PR #833 wired the per-request AbortSignal to cleanup for the two
invalidation-event SSE streams — but two more SSE routes share the
same retain shape and were missed:

• packages/server/src/gateway/routes/public/agent.ts — the agent SSE channel at GET /api/v1/agents/:agentId/events
• packages/server/src/gateway/gateway/index.ts — the worker SSE channel at GET /worker/stream

Both use Hono's streamSSE / stream helpers. Reading node_modules/hono/dist/helper/streaming/sse.js, Hono only wires signal -> stream.abort() inside an isOldBunVersion() branch. On Node + current Bun the only thing that triggers onAbort subscribers is ReadableStream.cancel() — which, under LB idle timeout / proxy kill / client hard close, does not fire.

Concretely, under abnormal disconnect on each of these routes:

• the 30s heartbeat setInterval keeps firing forever,
• SseManager (agent stream) keeps the dead connection in its per-agent Set,
• WorkerConnectionManager (worker stream) keeps the dead writer until the 10-minute stale-cleanup sweeps it,
• the while (!stream.aborted) await stream.sleep(1000) loop never exits.

All three retain the closure over the dead stream — the same leak shape #833 fixed for invalidation streams. Credible contributor to the OOMKilled-at-1Gi event tracked in #782.

Fix

New helper packages/server/src/events/sse-abort-bridge.ts centralizes the abort bridge so the pattern doesn't drift again the next time we add an SSE route. Both Hono-streaming routes now:

1. Bail early if requestSignal.aborted is already true at handler entry.
2. Wire stream.onAbort to local cleanup (interval clear + map evict).
3. Bridge requestSignal -> stream.abort() via the helper, so the onAbort subscribers run on abnormal disconnects too.
4. detach() the signal listener in a finally so the natural exit path doesn't leak it.

Reproducer

Added src/__tests__/unit/sse-abort-bridge.test.ts with a Hono streamSSE-mounted integration test:

• Red (no bridge — verified locally with the production code path before applying the fix): after the per-request AbortController.abort() fires, the heartbeat setInterval is still active and the while-!aborted loop has not exited within a 500ms grace window. Log line from the control run: OLD reproduce: loopExited = false timers = 1.
• Green (with bridge): loop exits within ~25ms and activeTimers.size returns to 0.

The control test was kept locally to validate the failure mode and then deleted from the repo to keep the committed suite tight; the green assertion is what's tested in CI.

Test plan

• bun test src/__tests__/unit/sse-abort-bridge.test.ts -> 8/8 pass
• bun test src/__tests__/unit/sse-cleanup.test.ts -> 5/5 pass (the existing fix(server): tear down SSE keepalive + listener on abnormal disconnect #833 suite, unchanged)
• bun run typecheck -> clean
• make build-packages -> clean
• Production validation: same as fix(server): tear down SSE keepalive + listener on abnormal disconnect #833 — monitor summaries-app-lobu-app RSS slope over 24h post-deploy and confirm it flattens, or grab a heap-snapshot diff via the flags plumbed in lobu-ai/owletto-web#140.

Scope

SSE-listener teardown only. The broader heap-leak hunt in #782 stays open until prod confirms. Other prime suspects from #782 not touched here:

• recordLifecycleEvent / metric_series train — bounded by inspection (single INSERT, 5s statement_timeout, 2000-row cap; no accumulating state on the call path).
• chat-state-adapter — no such file in the codebase; the path lives in the chat-SDK package and uses Postgres-backed state (chat_state_lists).
• pino redaction — the codebase uses winston, not pino.

[coderabbitai]

Warning

Rate limit exceeded

@buremba has exceeded the limit for the number of commits that can be reviewed per hour. Please wait 4 seconds before requesting another review.

You’ve run out of usage credits. Purchase more in the billing tab.

⌛ How to resolve this issue?

After the wait time has elapsed, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

We recommend that you space out your commits to avoid hitting the rate limit.

🚦 How do rate limits work?

CodeRabbit enforces hourly rate limits for each developer per organization.

Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout.

Please see our FAQ for further information.

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 21d907f5-3138-483d-9499-17ea8c8432a6

📥 Commits

Reviewing files that changed from the base of the PR and between de4c238 and d5aef8b.

📒 Files selected for processing (4)

• packages/server/src/__tests__/unit/sse-abort-bridge.test.ts
• packages/server/src/events/sse-abort-bridge.ts
• packages/server/src/gateway/gateway/index.ts
• packages/server/src/gateway/routes/public/agent.ts

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch chore/mem-leak-hunt-782

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!