# Event Manager Pipeline Report

- Generated: 2026-03-13T10:37:57.881502+00:00
- Session ID: `golden-summit-north`
- Overall status: **ok**
- Selected stages: meeting, salzburg, expo

## Stage results
- meeting:scrape: ok (1781.71s)
- salzburg:scrape: ok (118.65s)
- expo:scrape: ok (38.84s)
- meeting:deep_dive: ok (18.42s)
- salzburg:deep_dive: ok (13.19s)
- expo:deep_dive: ok (7.25s)
- meeting:manual_phase4: ok (6.98s)
- salzburg:manual_phase4: ok (145.99s)
- expo:manual_phase4: ok (6.82s)

## Phase 4 (human-like manual browsing) results
- meeting.vienna.info: checked 0, recovered_email 0, recovered_form_only 0, still_missing 0, blocked 0
- salzburgcongress.at: checked 12, recovered_email 0, recovered_form_only 1, still_missing 2, blocked 2
- expo-experts.at: checked 0, recovered_email 0, recovered_form_only 0, still_missing 0, blocked 0

## What went well?
- Pipeline executed with stage-by-stage telemetry and persisted logs.
- Sequential orchestration stayed stable across all stages.
- Reporting artifact generated and indexed under `/reports`.

## What went wrong?
- Data-quality issues can still appear from upstream link anomalies and recurring edge cases.
- Stage duration drift may occur; no strict timeout/escalation policy yet.

## Lessons learned
- Keep both ingestion-side sanitization and UI-side normalization for resilience.
- Post-run reporting improves observability and speeds up debugging.
- Structured logs are critical for identifying regressions early.

## Did we achieve the goal?
- **Yes**: run completion target met for session `golden-summit-north`.

## How can we improve?
- Add per-stage timeout and explicit timeout status in logs.
- Add collision trend metrics (last 7 runs) to detect regressions.
- Auto-link malformed outbound URLs to a normalized URL helper in UI.
- Add optional Slack/Telegram compact status card after each daily run.
- Add automatic report generation at end of every pipeline run.

## Small improvement plan (bird's-eye view)
- **Reliability:** add per-stage timeout + retry policy (capped retries).
- **Data quality:** add collision trend checks (7-run rolling window).
- **Observability:** auto-generate report after every full run.
- **UX:** keep `/settings -> /reports` path prominent and stable.
- **Operations:** add compact Telegram run-card with key metrics.

## References
- Pipeline log: `/home/clawdbot/clawd/Event_management/logs/event-manager-pipeline.jsonl`
- Meeting log: `/home/clawdbot/clawd/Event_management/logs/meeting-vienna-info.jsonl`
- Salzburg log: `/home/clawdbot/clawd/Event_management/logs/salzburgcongress.jsonl`
- Expo log: `/home/clawdbot/clawd/Event_management/logs/expo-experts.jsonl`
- Meeting Phase4 stream: `/home/clawdbot/clawd/Event_management/logs/meeting-vienna-info-manual-phase4.ndjson`
- Salzburg Phase4 stream: `/home/clawdbot/clawd/Event_management/logs/salzburgcongress-manual-phase4.ndjson`
- Expo Phase4 stream: `/home/clawdbot/clawd/Event_management/logs/expo-experts-manual-phase4.ndjson`