---

name: observability-review description: Evaluate logging coverage, error tracking, debugging capabilities, and monitoring patterns. Use when investigating production issues becomes difficult, after incidents, or when expanding monitoring coverage.

Observability Review

Assess the ability to understand, debug, and monitor the application in production.

Scope

1. Logging Coverage (per AGENTS.md)

• Structured logs with context (user id, request id)
• Sensitive data redacted before logging
• Appropriate log levels (error, warn, info, debug)
• Server actions logging mutations
• No excessive logging impacting performance

2. Error Tracking (Sentry)

• Sentry configured per AGENTS.md
• Performance traces for slow APIs
• Breadcrumb data for client errors
• Error boundaries on critical routes
• Source maps uploaded for stack traces
• User context attached to errors

3. Analytics (PostHog)

• Event tracking via lib/posthog/client.ts (client) and lib/posthog/server.ts (server)
• Feature flags with UPPERCASE_WITH_UNDERSCORE naming
• Custom properties in enums/const objects when reused
• No API key hallucination (use .env values)
• Key user actions tracked

4. Activity System (per CLAUDE.md)

• Activity logging via lib/activity/events/
• Domain-specific handlers (tasks, projects, clients, etc.)
• Overview cache (activity_overview_cache) maintained
• Audit trail for compliance

5. Health Monitoring

• API endpoint health checks
• Database connection monitoring
• External service health (Supabase, Resend)
• Queue processing status (Supabase Queues)

6. Debugging Capabilities

• Can reproduce issues from error reports?
• Request tracing from client to database?
• State inspection available?
• Time-based correlation possible?

7. Alerting Readiness

• Critical errors trigger notifications?
• Performance degradation detectable?
• Failed background jobs visible?
• Security events flagged?

Observability Gaps Analysis

For each area, identify:

• What we CAN see today
• What we CANNOT see
• Impact of blind spots
• Recommended additions

Output Format

Copy[PILLAR: LOGGING|ERRORS|METRICS|TRACES|ANALYTICS]
[PRIORITY: CRITICAL|HIGH|MEDIUM|LOW]
Area: What's being assessed
Current State: What exists
Gap: What's missing
Impact: How this affects debugging/monitoring
Recommendation: What to add
Effort: LOW|MEDIUM|HIGH

Actions

1. Inventory current logging statements
2. Review Sentry configuration and coverage
3. Audit PostHog event tracking
4. Check activity logging completeness
5. Identify production debugging pain points

Observability Patterns (from AGENTS.md)

Copy// Structured logging with context
console.log(JSON.stringify({
  level: 'info',
  action: 'task_created',
  userId: user.id,
  taskId: task.id,
  timestamp: new Date().toISOString()
}))

// Activity event logging
import { logTaskCreated } from '@/lib/activity/events/tasks'
await logTaskCreated(taskId, userId)

// PostHog server-side tracking
import { posthogServer } from '@/lib/posthog/server'
posthogServer.capture({
  distinctId: userId,
  event: 'task_completed',
  properties: { taskId, projectId }
})

Key Questions

1. When a user reports a bug: Can we find their session and see what happened?
2. When an API is slow: Can we identify the bottleneck?
3. When data looks wrong: Can we trace how it got that way?
4. When a deploy breaks something: Can we correlate timing with errors?
5. When we need usage patterns: Can we answer product questions?

Post-Review

Generate:

• Observability coverage map
• Blind spot inventory
• Recommended instrumentation additions
• Alerting rule suggestions
• Dashboard recommendations for Sentry/PostHog