cloudthinker-ai Event-Driven Architecture Review Skill
Review event-driven architecture for {{ system_name }} on {{ event_platform }}.
Workflow
Phase 1 — Event Inventory
EVENT CATALOG
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[ ] Total event types: ___
[ ] Event map:
Event Name | Producer | Consumer(s) | Volume/day
____________________|______________|______________|___________
| | |
| | |
| | |
[ ] Event schema registry: [ ] YES [ ] NO
[ ] Schema format: [ ] Avro [ ] JSON Schema [ ] Protobuf [ ] None
[ ] Event naming convention: ___
Phase 2 — Event Design Review
EVENT SCHEMA QUALITY
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[ ] Events contain:
[ ] Unique event ID
[ ] Event type/name
[ ] Timestamp (ISO 8601)
[ ] Source/producer identifier
[ ] Correlation/trace ID
[ ] Schema version
[ ] Events are self-contained (no need to fetch additional data)
[ ] Events represent facts (past tense: OrderPlaced, not PlaceOrder)
[ ] Event payload size within limits (avg: ___KB, max: ___KB)
[ ] No PII in events (or encrypted if required)
SCHEMA EVOLUTION
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[ ] Backward compatibility enforced
[ ] Forward compatibility considered
[ ] Schema versioning strategy:
[ ] Full compatibility
[ ] Backward only
[ ] None (breaking changes allowed)
[ ] Schema validation on publish: [ ] YES [ ] NO
Phase 3 — Reliability Patterns
RELIABILITY
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Delivery guarantees:
[ ] At-least-once delivery implemented
[ ] Consumer idempotency:
- Idempotency key strategy: ___
- Deduplication window: ___
[ ] Ordering guarantees:
- Partition/ordering key: ___
- Per-key ordering: [ ] GUARANTEED [ ] BEST EFFORT
[ ] Dead letter queue (DLQ) configured:
- DLQ destination: ___
- DLQ alert threshold: ___
- DLQ reprocessing procedure: [ ] DOCUMENTED
ERROR HANDLING MATRIX
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Error Type | Strategy | Max Retries | Backoff
Transient (network) | Retry | ___ | ___
Validation error | DLQ | 0 | N/A
Poison message | DLQ + alert | ___ | ___
Consumer crash | Rebalance | N/A | N/A
Phase 4 — Performance and Scalability
PERFORMANCE
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[ ] Throughput:
- Current: ___ events/sec
- Peak: ___ events/sec
- Target capacity: ___ events/sec
[ ] Latency (end-to-end, publish to consume):
- P50: ___ms
- P95: ___ms
- P99: ___ms
[ ] Consumer lag:
- Current lag: ___ events
- Lag alert threshold: ___ events
[ ] Partitions/shards: ___
[ ] Consumer group instances: ___
[ ] Backpressure handling: ___
Phase 5 — Observability
OBSERVABILITY
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[ ] Distributed tracing across events:
- Trace ID propagation: [ ] YES
- End-to-end trace visualization: [ ] YES
[ ] Metrics collected:
[ ] Publish rate per event type
[ ] Consumer processing rate
[ ] Consumer lag per partition
[ ] Error rate per consumer
[ ] DLQ depth
[ ] Alerting configured:
[ ] Consumer lag exceeds threshold
[ ] DLQ messages accumulating
[ ] Publisher failures
[ ] Consumer group rebalances
[ ] Event flow visualization/documentation: [ ] YES [ ] NO
Counter-Rationalizations
| Shortcut | Counter | Why |
|---|---|---|
| "We can skip some steps for this case" | Adapt the workflow steps, don't skip them | Skipped steps are where incidents and oversights originate |
| "The user seems to already know what to do" | Complete all workflow phases with the user | The workflow catches blind spots that experience alone misses |
| "This is a minor case, full process is overkill" | Scale the process down, don't turn it off | Minor cases become major when unstructured; the process scales, not disappears |
| "I'll fill in the details later" | Complete each section before moving on | Deferred details are forgotten; real-time capture is more accurate |
| "The template output isn't necessary" | Always produce the structured output format | Structured output enables comparison, audit trails, and handoff to other teams |
Output Format
Produce an event-driven architecture review report with:
- Architecture overview (event catalog, producers, consumers)
- Design quality (schema quality, naming, evolution strategy)
- Reliability assessment (delivery guarantees, error handling, DLQ)
- Performance profile (throughput, latency, scalability headroom)
- Recommendations (improvements ranked by impact and effort)