aws-samples / aws-samples/sample-autonomous-cloud-coding-agents
feat(orchestration): orchestrator DLQ and replay
- Dominant language
- TypeScript
- Stars
- 143
- Forks
- 46
- Avg merge
- 3d 9h
- Merged PRs (30d)
- 20
Description
**Context:** ROADMAP.md → Orchestrator DLQ
---
## Component
API or orchestration
## Describe the feature
**Dead-letter path** for task orchestration after durable execution retry exhaustion. Operators inspect and **replay** failed orchestration runs.
## Use case
Transient AWS faults or code bugs can exhaust durable retries with no durable artifact for ops. Webhook processor has DLQ (#284); orchestrator does not.
## Proposed solution
1. Configure DLQ on durable orchestrator Lambda (SQS).
2. Payload: task_id, checkpoint, error, timestamp.
3. CloudWatch alarm on DLQ depth (#117 pattern).
4. Operator API/CLI: list DLQ messages, replay with idempotency guard.
5. Document runbook in `docs/design/ORCHESTRATOR.md`.
## Other information
- Distinct from FanOutConsumer DLQ (#117).
- Design context: `docs/design/ORCHESTRATOR.md`.
- [ ] This might be a breaking change
Contributor guide
Research direction
Start with ROADMAP.md and docs/design/ORCHESTRATOR.md, then compare the webhook processor DLQ in #284 and the CloudWatch alarm pattern in #117. Define the DLQ, payload, alarm, operator API/CLI listing and replay flow, idempotency guard, and runbook; done means these pieces are implemented and documented.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- aws, typescript
- Domain
- backend, cli, cloud, distributed-systems, observability
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100