Files
grpc-library/docs/runbooks/grpc-platform-operations.md
T
DongHyeonka e6db162932 fix!: 재리뷰가 찾은 결함과 색인에서 빠졌던 발견 5건을 고친다
문제: 성숙도 리뷰 수정을 다시 읽은 재리뷰(Critical 0, 회귀 0)가
Important 27건·Minor 41건을 찾았다. 수정이 새로 만든 결함이 많았다.
@SystemInvoked가 권한 규칙의 탈출구가 됐고, 데모 계정 거부가 정확히
prod일 때만 걸렸고, outbox 순서 번호가 업그레이드 때 뒤로 갔고,
GraphQL union 아래 복잡도가 0이었고, 서블릿 오류 응답이 success:true로
나갔고, 호출자 deadline이 breaker를 열었다. 원래 리뷰 색인이 06 영역
발견 5건(06-10~14)을 빠뜨려 그동안 처리되지 않았다.

방법: 영역별 7묶음이 test-driven-development 절차로 실패 테스트를
먼저 쓰고 고쳤다. 결정은 재리뷰 README 4절에 적었다: 순서 번호는 V21이
기록한 전역 floor부터, 데모 계정은 local·test·demo 허용 목록,
@SystemInvoked bean은 기동 실패, complete 단계 실패는
IDEMPOTENCY_COMPLETION_UNKNOWN(결과 불명), 캐시 조건부 쓰기는 별도
SCRIPT lane, strict lane 정책은 build cache 입력, 끝나지 않는 재시도
행은 원인 코드·지표·PARKED 절차. 레지스트리가 선언만 하고 아무 코드도
내보내지 않던 redis.capability.* 지표 6개는 빼고 runbook을 실제
지표(redis.lane.*) 기준으로 고쳤다.

위험·한계: V21을 V22 없이 직접 고쳤다(어느 배포에도 아직 없음). 이
브랜치의 V21을 적용한 개발 DB는 다시 만들어야 한다. 서블릿 오류
디스패치를 보안 체인이 허용한다. 발행·저장소 포트 시그니처가 바뀌어
파생 서비스 코드가 깨진다. prod에서 fixture-로 시작하는 설정값은
기동을 막는다. httpclient failure-injection lane의 main 단계 중복
실행은 이전 결정대로 남겼다. 실제 CI runner 실행은 확인하지 못했다.

검증: ci, mainCheck, :integrationCheck, -p optional-platforms ci
integrationCheck performanceCheck, :performanceCheck, :releaseCheck
(-PgitRevision), verify-compose-profile-contracts(18 lane) 통과.
첫 실행의 실패 2건(FindSecBugs 1, sample 동시 생성 테스트의 v4 event
id)은 고친 뒤 ci·integrationCheck·mainCheck를 다시 돌려 통과했다.

Refs: docs/reviews/2026-11-03-rereview/README.md 3·4·6·7절
BREAKING CHANGE: outbox 발행 포트와 저장소 포트 시그니처, 순서 번호
첫 값(V21 floor 다음), 샘플 데모 계정 프로필 허용 목록, 새 오류 코드
IDEMPOTENCY_COMPLETION_UNKNOWN, prod의 fixture- 값 거부, 기동 rule은
static·의존 없음이어야 한다.
2026-11-03 23:29:35 +09:00

9.1 KiB

Runbook: gRPC platform operations

Scope: the :grpc:* family. All of it is build-only today — every leaf's runtime_memberships is empty — so nothing here fires in production yet. It is written now because the states it covers are the ones an on-call cannot work out from first principles at three in the morning, and shipping the behaviour before the runbook means the first person to meet one is doing that.

Configuration lives under app.grpc.platform.* and is bound by GrpcPlatformProperties. The platform does not start unless app.grpc.platform.enabled=true.


COMPLETION_UNKNOWN on a mutation

What you are seeing. A client received DEADLINE_EXCEEDED, UNAVAILABLE or INTERNAL on a state-changing call, and the response trailer completion-outcome reads COMPLETION_UNKNOWN.

What it means. The server may have committed. This is not a failure and not a success; the status code cannot distinguish them, which is why the outcome is carried separately.

What not to do. Do not re-issue the call. Do not tell the caller it failed. Both are wrong half the time, and which half is not knowable from the status.

What to do.

  1. Take the error-execution-id from the trailers. It is the only link between what the client saw and what the server did.

  2. If the method is IDEMPOTENCY_KEY_REQUIRED, query the operation ledger with the caller fingerprint, the full method name and the caller's key. GrpcOperationStatusQuery returns one of IN_PROGRESS, COMMITTED, FAILED_TERMINAL, NOT_FOUND or UNKNOWN.

  3. COMMITTED means return the stored outcome reference, not a freshly computed answer — the resource may have changed since, and a new answer would describe the state at reconciliation time rather than the state the caller's own call produced.

  4. NOT_FOUND means the operation never started and is safe to re-issue. FAILED_TERMINAL means the same.

  5. UNKNOWN means the ledger could not be consulted. Nothing may be concluded. The case is queued by GrpcCompletionReconciler and retried later.

  6. If the method is not keyed, there is no ledger row. Resolve it against the business resource, or escalate to the service owner. This is the case the keyed profile exists to avoid.

  7. IN_PROGRESS means a claim exists and has not finished. If its owner died, the claim is not stuck forever: once it is older than the claim TTL (app.grpc.platform.operation-ledger-claim-ttl, default 5m), the next retry of the same request takes it over and runs again, and a later reconciler pass sees that attempt's COMMITTED or FAILED_TERMINAL. Until a retry arrives, the case stays pending. The re-run is safe only when the ledger commit is recorded in the mutation's own transaction; if this service commits them separately, check the business resource before letting the retry through. An owner that was only slow cannot spoil the takeover: its claim attempt (claim_attempt) is behind, so its late commit is refused and its late failure ignored. A log line saying an operation "is no longer held by attempt N" is that refusal, not an outage.

Escalate when the reconciler's pending list grows across passes. That means the ledger is unreachable rather than slow. A case pending as IN_PROGRESS for longer than the claim TTL with no retry arriving is a caller that gave up; it needs the business resource checked, not a ledger fix.


The operation ledger table keeps growing

What you are seeing. grpc_operation_ledger grows without bound, or a retry of an old key runs the mutation again instead of replaying.

What it means. Finished rows (COMMITTED, FAILED_TERMINAL) are deleted only when something calls GrpcOperationLedgerRetention.purgeExpired(now); the starter assembles that bean (app.grpc.platform.operation-ledger-enabled=true) but does not schedule it. A growing table means the service's scheduler is not calling it. A retry that re-ran after the retention (app.grpc.platform.operation-ledger-retention, default 24h) is the trade the retention makes: its record was purged, so it was a new operation.

What to do.

  1. Check that the service schedules purgeExpired. Each call deletes in batches of 500 until a batch comes back short, never touching IN_PROGRESS rows.
  2. For a large backlog, call it repeatedly; each batch is a short transaction that skips rows another transaction holds.
  3. If retried keys re-run after the retention, raise the retention above the callers' retry window rather than skipping the purge.

Escalate when old IN_PROGRESS rows accumulate. Those are not purged by design; each is an abandoned claim no retry has taken over (see IN_PROGRESS above).


A stream ended with FULL_RESYNC_REQUIRED

What you are seeing. A client's resume was refused and it was told to resynchronise.

What it means. The server can no longer replay from the client's cursor. Either the snapshot version moved, or the cursor predates retained history.

What to do. Nothing on the server. The client is expected to discard its position and start a new stream from a fresh snapshot. A client that instead retries the same token will keep receiving the same answer.

Escalate when it is happening to many clients at once. That usually means history retention was reduced, or snapshots are rotating faster than clients reconnect.


A stream ended with SLOW_CONSUMER

What you are seeing. Streams terminating with SLOW_CONSUMER, and grpc.stream.flow_control_stalls rising.

What it means. The consumer could not keep up with the bounded queue. The stream was terminated rather than silently dropping messages, because a client cannot detect drops — the sequence numbers it sees are the ones it was sent.

What to do.

  1. Check whether the consumer is slow or the producer is fast. grpc.stream.messages against grpc.stream.lifetime tells you the rate.
  2. If the consumer is slow, the fix is on the consumer. Raising the queue bound moves the failure later and makes it larger.
  3. A resume is not available after this ending: the messages that overflowed the queue are gone, so continuing from the last delivered sequence would silently skip them. The client resynchronises.

RESOURCE_EXHAUSTED under load

What you are seeing. Calls refused with RESOURCE_EXHAUSTED and GrpcAdmissionController reporting rejections.

What it means. The server is at its concurrency and queue bounds and is shedding rather than queueing. This is the designed behaviour: accepting work whose callers have already given up spends capacity on nothing.

What to do.

  1. Read grpc.rpc.duration and grpc.rpc.queue_wait separately. Queue time rising with duration flat means the bottleneck is admission, not the work.
  2. Check which saturation counter is moving — executor, channel or flow control. They look identical in a latency graph and have different fixes.
  3. Raising app.grpc.platform.executor-queue-capacity defers the problem; it does not remove it. GrpcExecutorProfile refuses a queue above ten thousand for that reason.

A rollout is producing errors at every deploy

What you are seeing. A burst of UNAVAILABLE or CANCELLED each time an instance goes away.

What it means. The drain sequence is not completing, or is running out of order.

What to do.

  1. GrpcDrainResult records what each drain achieved: completed and cancelled unary calls, signalled and cancelled streams, and which phases ran. A drain that routinely force-cancels is the cause.
  2. The order matters. Readiness flips first and nothing is refused during that window, because there is a gap between an instance reporting unready and routing acting on it. Refusing during that gap turns a clean rollout into a burst of errors at every deploy.
  3. For long streams, check the Kubernetes profile's streamReconnectBudget and readinessDrainGrace. A stream is pinned to one pod for its whole life, so every rollout ends it; a profile with long streams and no reconnect budget has not decided what clients do next.

Verifying a deployment's configuration

GrpcPlatformSnapshotService produces a secret-free snapshot for a caller on the admin network holding an admin role. Both gates are required.

GrpcPlatformSnapshotService.driftAgainstRelease compares a running snapshot with the release manifest and reports schema version, method policy hash and per-channel profile differences. An instance running a configuration the release did not ship is behind a whole class of incidents that are otherwise diagnosed by reading logs.

The snapshot carries hashes and names only. A field whose name looks like a credential is refused at construction rather than redacted.


Things that are deliberately off

  • Reflection in production. GrpcReflectionMode.defaultFor returns DISABLED for STAGE and PROD. Reflection publishes the whole schema to anyone who can open a connection.
  • Every advanced capability. See docs/runbooks/grpc-advanced-capabilities.md.
  • The platform itself. app.grpc.platform.enabled defaults to false, and every :grpc:* leaf is build-only in the registry.