cloudthinker-ai Cache Flush and Warm Runbook Skill
Execute cache flush for {{ cache_service }} at {{ cache_endpoint }}. Scope: {{ flush_scope }} | Reason: {{ reason }}
Workflow
Phase 1 — Impact Assessment
IMPACT ASSESSMENT
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
CACHE PROFILE
Service: {{ cache_service }}
Endpoint: {{ cache_endpoint }}
Flush scope: {{ flush_scope }}
Current memory usage: ___ MB / ___ MB total
Key count: ___
Hit ratio (current): ___%
DOWNSTREAM IMPACT
[ ] Identify services dependent on this cache
[ ] Estimate cache-miss penalty (latency increase): ___ ms
[ ] Estimate origin load increase: ___x multiplier
[ ] Check origin (database/API) can handle thundering herd
[ ] Estimate time to rebuild cache organically: ___ minutes
RISK ASSESSMENT
[ ] Can origin handle 100% cache-miss traffic? YES / NO
[ ] Is there a read-through cache pattern in place? YES / NO
[ ] Will flush cause user-visible latency increase? YES / NO
[ ] Is there a risk of cache stampede? YES / NO
Phase 2 — Pre-Flush Preparation
PREPARATION
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[ ] Notify dependent service owners
[ ] Scale up origin (database / API) if needed
- Current capacity: ___
- Target capacity: ___
[ ] Enable rate limiting on cache-miss path (if available)
[ ] Configure circuit breaker thresholds (if available)
[ ] Take cache snapshot / RDB backup (if full flush)
[ ] Open monitoring dashboards:
- Cache hit/miss ratio
- Origin service latency and error rate
- Application response times
[ ] Confirm maintenance window with stakeholders
Phase 3 — Flush Execution
FLUSH EXECUTION
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
TARGETED FLUSH (preferred — specific keys/patterns):
1. [ ] Identify keys to flush: {{ flush_scope }}
2. [ ] Count keys matching pattern: ___
3. [ ] Delete keys using SCAN + DEL (never use KEYS in production)
4. [ ] Verify target keys removed
5. [ ] Confirm non-target keys intact
FULL FLUSH (use only when necessary):
1. [ ] Final confirmation: full flush approved by ___
2. [ ] Execute FLUSHDB or FLUSHALL
3. [ ] Record flush timestamp: ___
4. [ ] Verify memory freed: ___ MB
POST-FLUSH METRICS
Key count after flush: ___
Memory after flush: ___ MB
Timestamp: ___
Phase 4 — Cache Warming Strategy
CACHE WARMING
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
WARMING APPROACH (select one):
[ ] Passive warming — let read-through rebuild cache naturally
[ ] Active warming — run warming script to pre-populate hot keys
[ ] Hybrid — warm critical keys actively, let others rebuild passively
ACTIVE WARMING STEPS (if selected):
1. [ ] Identify hot keys from previous access patterns
2. [ ] Run warming script / job:
- Target key count: ___
- Batch size: ___
- Rate limit: ___ keys/second
3. [ ] Monitor warming progress: ___% complete
4. [ ] Verify warmed data correctness (spot-check samples)
WARMING PROGRESS
Hit ratio target: > ___%
Current hit ratio: ___%
Estimated time to target: ___ minutes
Phase 5 — Post-Flush Validation
POST-FLUSH VALIDATION
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
CACHE HEALTH (check at T+5min, T+15min, T+1h)
[ ] Cache hit ratio recovering: ___% -> ___% -> ___%
[ ] Memory utilization trending back to normal
[ ] No eviction pressure (if memory-limited)
[ ] Connection count stable
APPLICATION HEALTH
[ ] Response times returning to baseline
[ ] Error rate at or below pre-flush level
[ ] No timeout errors from cache layer
[ ] Origin load decreasing as cache warms
ORIGIN HEALTH
[ ] Database CPU / connections returning to normal
[ ] API backend latency returning to normal
[ ] No connection pool exhaustion events
Phase 6 — Cleanup
CLEANUP
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[ ] Scale down origin if it was scaled up
[ ] Remove temporary rate limits or circuit breaker overrides
[ ] Confirm cache hit ratio at steady state: ___%
[ ] Notify stakeholders of completion
[ ] Document flush results and lessons learned
[ ] Update cache TTL policies if stale data was the root cause
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 a cache flush execution report with:
- Flush summary (service, scope, reason, timestamps)
- Impact assessment results and risk mitigations applied
- Flush execution confirmation with before/after metrics
- Warming progress tracking and hit ratio recovery curve
- Validation results (application and origin health)