Phase 72.5.2.8 adds @v2/crash-analytics-dashboard, the operational crash
analytics surface for crash rate by version, platform, hardware, and top crash
stacks. The policy is v2-crash-analytics-dashboard-v1.
buildV2CrashAnalyticsDashboardSurface anchors to @tara/monitoring error
event vocabulary and aggregates V2 v2.player.session.started plus
v2.match.crash.reported telemetry. It emits chart-ready crash-rate rows for
version, platform, and hardware tier, plus top crash stacks grouped by
normalized stack signature.
The service consumes accountIdHash only as hashed-account telemetry and never
emits raw account IDs. Raw stack traces are accepted only as an input to produce
stable stack signatures and normalized top-frame labels; the surface does not
expose raw stack traces.
Operator Surface#
Rendered UI should lead with crash rate, crash-free session rate, and crash event volume. Keep version, platform, hardware, and date filters compact and shared across every table.
The first screen should make it clear which release version, platform, and hardware tier are producing the highest crash rate. Top crash stack rows should show stack signature, normalized top frame, affected sessions, first/last seen times, and whether symbolication coverage is present. Avoid marketing copy and show empty states when crash telemetry has not arrived.
Verification#
Targeted verification:
pnpm --filter @v2/crash-analytics-dashboard run typecheck
pnpm --filter @v2/crash-analytics-dashboard run test
python V2/ue/Tools/check-v2-crash-analytics-dashboard.py
python -m json.tool V2/ue/Build/Telemetry/v2-crash-analytics-dashboard.json