## Phase 149: Sovereign Crash Analytics & Error Tracking Platform — Sentry / BugSnag / Rollbar / Crashlytics / Raygun / Honeybadger / Datadog Error Tracking Replacement

Closes the error-tracking sovereignty gap. Existing phases reference Sentry as a
dependency across services (Phase 3, Phase 22, Phase 30–33, others). Phase 98
covers AI observability. Phase 145 CI/CD collects build telemetry. None of these
deliver a **dedicated multi-language, multi-platform crash ingestion,
symbolication, grouping, release-health, and distributed-tracing platform**
capable of replacing Sentry across Oshun's 45+ domain apps.

This phase delivers a complete sovereign replacement for Sentry, BugSnag,
Rollbar, Firebase Crashlytics, Raygun, Honeybadger, Airbrake, and the
error-tracking portions of Datadog / New Relic / Instabug.

---

### 149.1 @neith/crash-ingestion — Event Ingestion & SDK

- [x] **149.1.1** JavaScript / TypeScript browser SDK (source-map aware)
- [x] **149.1.2** Node.js server SDK with Express / Fastify / Nest middleware
- [x] **149.1.3** iOS / macOS Swift SDK with crash-reporter bridge
- [x] **149.1.4** Android Kotlin / Java SDK with NDK native bridge
- [x] **149.1.5** Rust SDK with backtrace + panic hook
- [x] **149.1.6** Go SDK with panic / error wrapping
- [x] **149.1.7** Python SDK with Sentry-API-compat surface
- [x] **149.1.8** C / C++ SDK with breakpad / crashpad integration
- [x] **149.1.9** Unity / Neith Engine plugin
- [x] **149.1.10** Unreal / Neith Engine plugin
- [x] **149.1.11** Flutter / Dart SDK
- [x] **149.1.12** React Native / Expo SDK
- [x] **149.1.13** Electron main + renderer SDK
- [x] **149.1.14** Kubernetes sidecar ingest agent
- [x] **149.1.15** OTLP bridge (accept Sentry-envelope + OTLP traces)
- [x] **149.1.16** Sentry DSN-compat endpoint for drop-in replacement
- [x] **149.1.17** Rate-limited client buffering with exponential backoff
- [x] **149.1.18** Offline queue with journalled replay
- [x] **149.1.19** PII scrubbing at SDK layer (client-side strip)
- [x] **149.1.20** SDK auto-update channel

---

### 149.2 @neith/crash-symbolicator — Symbolication & Deminification

- [x] **149.2.1** iOS dSYM symbolication
- [x] **149.2.2** Android ProGuard / R8 mapping deobfuscation
- [x] **149.2.3** Android NDK breakpad symbol upload
- [x] **149.2.4** JavaScript source-map resolver with chain walking
- [x] **149.2.5** Windows PDB symbolication
- [x] **149.2.6** Linux DWARF + debug-link symbol lookup
- [x] **149.2.7** Rust / Go / Swift native stack trace demangling
- [x] **149.2.8** WebAssembly name-section + DWARF
- [x] **149.2.9** Unity IL2CPP line.cpp translation
- [x] **149.2.10** Unreal VS / AS profile symbolication
- [x] **149.2.11** Symbol server (S3-compat) with CI auto-upload
- [x] **149.2.12** Symbol server retention by release cohort
- [x] **149.2.13** Reprocess events when late-uploaded symbols arrive
- [x] **149.2.14** Source-context lookup (show ±5 lines of code in trace)
- [x] **149.2.15** Git blame integration per frame (culprit commit)

---

### 149.3 @neith/crash-grouping — Fingerprinting, Grouping & Dedup

- [x] **149.3.1** Stack-trace fingerprint (normalized frames)
- [x] **149.3.2** Exception-type + message fuzzy grouping
- [x] **149.3.3** Hierarchical grouping (loose / strict / custom)
- [x] **149.3.4** User-defined fingerprint rules (regex / wildcard)
- [x] **149.3.5** ML-based grouping (embedding clustering)
- [x] **149.3.6** Split group (over-grouping correction)
- [x] **149.3.7** Merge groups (under-grouping correction)
- [x] **149.3.8** Event-to-issue assignment with atomic dedupe
- [x] **149.3.9** Issue state machine (unresolved / resolved / ignored /
      reopened)
- [x] **149.3.10** Regression detection (resolved → reopened)
- [x] **149.3.11** Auto-resolve when next release ships
- [x] **149.3.12** Snooze until next occurrence / count / time
- [x] **149.3.13** Issue merge / unmerge from UI
- [x] **149.3.14** Release-specific grouping scope
- [x] **149.3.15** Cross-project issue link (same bug across services)

---

### 149.4 @neith/crash-storage — Event Storage & Query

- [x] **149.4.1** ClickHouse-class column store for events (≥ 10M / day)
- [x] **149.4.2** S3-compat blob store for full event payload
- [x] **149.4.3** Retention by plan / project (7 / 30 / 90 / 365 d)
- [x] **149.4.4** Sampling at ingest with deterministic hash
- [x] **149.4.5** Search by user / release / tag / env / URL / custom
- [x] **149.4.6** Full-text search on message + trace
- [x] **149.4.7** Saved search with notification rule
- [x] **149.4.8** Query builder UI (drag-drop filters)
- [x] **149.4.9** SQL-like query API for advanced users
- [x] **149.4.10** Aggregate views (count per group per hour)

---

### 149.5 @neith/crash-releases — Release Health & Tracking

- [x] **149.5.1** Release entity with commit-range + version
- [x] **149.5.2** Deploy event with environment + timestamp
- [x] **149.5.3** Crash-free-session rate per release
- [x] **149.5.4** Crash-free-user rate per release
- [x] **149.5.5** Adoption graph (% of users on release)
- [x] **149.5.6** First-seen / last-seen per release per issue
- [x] **149.5.7** Regression-introduced-by release attribution
- [x] **149.5.8** Commit blame inside release diff
- [x] **149.5.9** Release comparison (A vs B delta)
- [x] **149.5.10** Automatic release creation from CI (Phase 145)
- [x] **149.5.11** Session tracking SDK (start / heartbeat / crash / exit)
- [x] **149.5.12** Session boundary rules (on background / after 30s idle)
- [x] **149.5.13** Release threshold alert (crash-rate > X %)
- [x] **149.5.14** Halt-deploy hook when threshold exceeded
- [x] **149.5.15** Rollback trigger via CI (Phase 145)

---

### 149.6 @neith/crash-sessions — Session Replay & User Context

- [x] **149.6.1** DOM recording (privacy-safe masked input)
- [x] **149.6.2** CSS + asset snapshot for faithful replay
- [x] **149.6.3** Console + network + click timeline
- [x] **149.6.4** Mobile screen recording (native view tree)
- [x] **149.6.5** Redaction rules (mask / block / custom selector)
- [x] **149.6.6** Replay rate-limit / sampling per session
- [x] **149.6.7** Replay player with scrubber + speed
- [x] **149.6.8** Replay tied to specific event / error ID
- [x] **149.6.9** Rage-click / dead-click / error-click detection
- [x] **149.6.10** Accessibility tree overlay in replay
- [x] **149.6.11** Network waterfall inside replay
- [x] **149.6.12** User feedback widget with screenshot
- [x] **149.6.13** Cookie / localStorage snapshot at error
- [x] **149.6.14** Replay retention by plan
- [x] **149.6.15** Replay export as MP4 for ticket attachment

---

### 149.7 @neith/crash-performance — Tracing & APM Bridge

- [x] **149.7.1** OTel-compatible distributed tracing ingest
- [x] **149.7.2** Span / transaction view with flamegraph
- [x] **149.7.3** N+1 query detection
- [x] **149.7.4** Slow-DB / slow-HTTP span classification
- [x] **149.7.5** Latency percentile (p50 / p75 / p95 / p99)
- [x] **149.7.6** Frontend Web Vitals (LCP / FID / CLS / INP / TTFB)
- [x] **149.7.7** Mobile startup trace (cold / warm / resume)
- [x] **149.7.8** Profile aggregation (sampled CPU profiles)
- [x] **149.7.9** Correlate trace → error → replay in one view
- [x] **149.7.10** Anomaly detection on latency regression

---

### 149.8 @neith/crash-alerting — Alerts, Notifications & Integrations

- [x] **149.8.1** Rule builder (issue / metric / frequency / threshold)
- [x] **149.8.2** Escalation policy integration (Phase 152)
- [x] **149.8.3** Digest summaries (daily / weekly)
- [x] **149.8.4** Slack / sovereign chat (Phase 139) destination
- [x] **149.8.5** Email / SMS / push destination
- [x] **149.8.6** Webhook destination with HMAC signing
- [x] **149.8.7** Issue-owner routing (by path / tag / team)
- [x] **149.8.8** Auto-assign from commit author / CODEOWNERS
- [x] **149.8.9** Noise suppression (spike protection)
- [x] **149.8.10** Alert cooldown + silence windows
- [x] **149.8.11** Jira / Linear / sovereign PM ticket create
- [x] **149.8.12** PR comment on related commits
- [x] **149.8.13** GitHub / sovereign forge (Phase 12) issue link
- [x] **149.8.14** Teams / Discord / Mattermost destination
- [x] **149.8.15** Mobile app for on-call with ack / assign

---

### 149.9 @neith/crash-security — PII, Compliance & Self-Host

- [x] **149.9.1** Server-side PII scrub (regex + named rules)
- [x] **149.9.2** Data scrubbing preview (see masked)
- [x] **149.9.3** Data residency per project (region pin)
- [x] **149.9.4** SOC 2 / HIPAA / GDPR audit controls
- [x] **149.9.5** Customer-managed keys (CMEK) encryption at rest
- [x] **149.9.6** Retention enforcement jobs
- [x] **149.9.7** Right-to-erasure GDPR API
- [x] **149.9.8** Consent-aware SDK (pause collection on opt-out)
- [x] **149.9.9** Self-host deployment profile (single-node / HA / regional)
- [x] **149.9.10** Air-gapped deployment option

---

### 149.10 @neith/crash-intelligence — AI-Assisted Triage

- [x] **149.10.1** LLM suggest-fix (Nous Phase 47) on stack trace
- [x] **149.10.2** Similar-issue clustering across tenants (opt-in)
- [x] **149.10.3** Auto-PR proposed fix for known patterns
- [x] **149.10.4** Root-cause narrative (LLM summary)
- [x] **149.10.5** Severity inference from user-impact + trace
- [x] **149.10.6** Known-noise classifier (auto-ignore / auto-resolve)
- [x] **149.10.7** Flaky-test vs real-bug discriminator
- [x] **149.10.8** Embedding search ("find issues like this one")
- [x] **149.10.9** Feedback loop fine-tunes triage model
- [x] **149.10.10** Suggested owner via code-graph + history

---

### 149.11 @neith/crash-ui — Dashboard & Reporting

- [x] **149.11.1** Issue feed with filters + saved views
- [x] **149.11.2** Issue detail (frames / context / breadcrumbs / replay)
- [x] **149.11.3** Trend graphs (events / users / sessions)
- [x] **149.11.4** Cross-project rollup dashboards
- [x] **149.11.5** Team-ownership dashboards
- [x] **149.11.6** Executive weekly report PDF
- [x] **149.11.7** SLO-style burndown of unresolved issues
- [x] **149.11.8** Command palette for power users
- [x] **149.11.9** Keyboard-first navigation
- [x] **149.11.10** Dark / high-contrast / reduced-motion a11y

---

### 149.12 Testing & Validation

- [x] **149.12.1** SDK integration test matrix (15 platforms)
- [x] **149.12.2** Symbolication fidelity vs Sentry reference ≥ 99 %
- [x] **149.12.3** Ingest throughput 10M events / day soak
- [x] **149.12.4** Grouping accuracy vs labelled corpus ≥ 95 %
- [x] **149.12.5** Replay size / fidelity benchmarks
- [x] **149.12.6** PII scrub regression tests
- [x] **149.12.7** Drop-in Sentry DSN compatibility conformance
- [ ] **149.12.8** Dogfood: serve every Oshun service (Phase 3–147)
      `blocked:upstream` Tagged 2026-09-18 against the range the item names,
      counted rather than assumed: 15 of the phase trackers between 3 and 147
      still hold open items, `TODOS/phase-27.md` alone 934. "Every Oshun
      service" is not a set that exists yet, and a dogfood over the subset that
      does would be a different claim from the one this box makes. The seven
      sibling gates in §149.12 — symbolication fidelity, ingest soak, grouping
      accuracy, replay benchmarks, PII scrub, DSN compatibility — are the ones
      measurable here, and all seven are checked.
