feat: event monitoring setup wizard + pipeline test tooling

[AhmadHammad21]

Summary

• 4-step setup wizard on /init — account creation → AWS config → permission check → infrastructure creation (SQS + 9 EventBridge rules + aggregate CloudWatch alarm)
• Settings → AWS Configuration — new Event Monitoring section with Create/Teardown infrastructure buttons and status indicator
• Monitoring page — Send test event button (admin only) for quick pipeline validation
• Backend — two new endpoints: DELETE /api/init/infra (teardown) and POST /api/init/test-event (synthetic SQS push)
• Aggregate CloudWatch alarm (opendevops-lambda-errors-aggregate) created automatically during setup — monitors all Lambdas account-wide, no per-function alarms needed
• Polling → /monitoring fix — _persist_and_notify now calls add_alert() so polling investigations appear on the Monitoring page
• Event consumer fix — service=unknown bug fixed, now reads from services_affected[0]
• LiteLLM log spam fix — suppresses "Provider List" info messages on startup
• Test script — scripts/test_pipeline.py (boto3-based, Windows-compatible) with --lambda-only flag for testing real alarm path vs direct SQS push
• Docs — docs/monitoring.md rewritten to cover both detection modes end-to-end; docs/slack_and_polling.md updated with comparison table

Test plan

Settings — Infrastructure (test these first, infra is live)

• Settings → AWS Configuration → Teardown deletes SQS queue, 9 EventBridge rules, and opendevops-lambda-errors-aggregate alarm in AWS console
• After teardown, UI shows "Disabled" badge and "Create Infrastructure" button
• Create Infrastructure button re-provisions everything and flips back to "Enabled"

Monitoring page

• Send test event button (admin) calls POST /api/init/test-event, shows toast, new card appears on /monitoring within ~90s
• Non-admin (user role) does not see the Send test event button
• Error toast shown when button is clicked but infra not set up (teardown first)

/init wizard (requires clean state — reset data/agent.db or use fresh env)

• Step 1 — admin account creation works, advances to Step 2
• Step 2 — region selector + optional SNS ARN, advances to Step 3
• Step 3 — permission check grid renders green ✓ / red ✗ badges per service; Required badge shown for cloudwatch/lambda/sqs/events
• Step 4 — "Create Infrastructure" creates SQS + rules + alarm, shows success summary
• Step 4 — "Skip for now" shows disclaimer listing 4 missing features; navigates to app
• Step indicator shows checkmarks for completed steps
• Navigating to /init when already initialized redirects away

Polling

• POLL_INTERVAL_MINUTES=1 + Lambda errors → investigation appears on /monitoring within ~2 min
• POLL_INTERVAL_MINUTES=0 (default) — poller never starts, no log lines about poller

Event-driven (direct SQS push — fastest)

• uv run python scripts/test_pipeline.py --function <name> → card on /monitoring in ~90s

Event-driven (real alarm path)

• uv run python scripts/test_pipeline.py --function <name> --lambda-only → alarm trips in ~2 min → card appears on /monitoring

Auth edge case

• Wizard works with JWT_SECRET unset (auth disabled mode)