| license | Apache-2.0 |
| name | mobile-analytics-crash-reporting-expert |
| description | Mobile analytics and crash reporting expert for Firebase, Sentry, Mixpanel, custom events, and crash symbolication. Activate on: mobile analytics, crash reporting, Firebase Analytics, Sentry mobile, Mixpanel, event tracking, crash symbolication, dSYM upload, ProGuard mapping, session replay. NOT for: server-side logging (use log-aggregation-architect), web analytics (use frontend-architect), A/B testing (use environment-config-manager). |
| allowed-tools | Read,Write,Edit,Bash(docker:*,kubectl:*,terraform:*,npm:*,npx:*) |
| category | Mobile Development |
| tags | ["analytics","crash-reporting","sentry","firebase"] |
| pairs-with | [{"skill":"logging-observability","reason":"Mobile observability extends the broader logging and observability strategy"},{"skill":"app-store-submission-automator","reason":"Crash-free rate metrics impact app store ranking"}] |
Mobile Analytics & Crash Reporting Expert
Expert in mobile analytics instrumentation, crash reporting with symbolication, and actionable event tracking across iOS and Android.
Decision Points
Stack Selection Decision Tree
High crash volume (>1000/day) OR complex error debugging needed?
├─ YES: Choose Sentry
│ ├─ Need user sessions? → Add Sentry Session Replay
│ └─ Enterprise budget? → Consider Bugsnag for enterprise features
│
└─ NO: Evaluate based on primary use case
├─ Need product analytics + basic crashes? → Firebase (free tier generous)
├─ Advanced funnel analysis required? → Mixpanel + Crashlytics
└─ Real-time user behavior insights? → Amplitude + Sentry
Budget considerations:
├─ <$50/month: Firebase Analytics + Crashlytics
├─ $50-500/month: Sentry Team + basic product analytics
└─ $500+/month: Full stack (Sentry + Mixpanel/Amplitude)
Sample Rate Tuning
Production traffic volume:
├─ <1K daily sessions:
│ ├─ Crash sampling: 100%
│ ├─ Transaction tracing: 50%
│ └─ Session replay: 25%
│
├─ 1K-10K daily sessions:
│ ├─ Crash sampling: 100%
│ ├─ Transaction tracing: 20%
│ └─ Session replay: 10%
│
└─ >10K daily sessions:
├─ Crash sampling: 100% (never sample crashes)
├─ Transaction tracing: 5-10%
└─ Session replay: 1-5%
Cost optimization triggers:
├─ Monthly bill >$200 → Reduce transaction sampling by 50%
├─ Storage quota exceeded → Lower session replay rate
└─ Performance impact detected → Disable replay in low-memory scenarios
Error Volume Handling
Incoming crash rate assessment:
├─ <0.1% of sessions → Normal monitoring, weekly review
├─ 0.1-1% of sessions → Daily triage, investigate top 3 crashes
├─ 1-5% of sessions → Emergency mode: stop releases, hotfix priority
└─ >5% of sessions → App store removal risk, immediate rollback
Symbolication status check:
├─ Symbolicated within 5 minutes → Continue normal flow
├─ Pending 5-30 minutes → Check CI upload job status
└─ Missing after 30 minutes → Manual dSYM/mapping upload required
Failure Modes
1. "Symbol Upload Amnesia"
Detection: Stack traces show memory addresses (0x1a2b3c4d) instead of function names
Diagnosis: dSYM/ProGuard mapping upload failed or incomplete
Fix:
- Check CI upload job logs for failures
- Manually upload:
sentry-cli upload-dif --org X --project Y ./dSYMs/
- Verify upload:
sentry-cli releases files <version> list
2. "PII Leak Catastrophe"
Detection: Analytics events contain email addresses, phone numbers, or names in Sentry dashboard
Diagnosis: Missing or broken data scrubbing rules
