Collapse the backend from 7 Coolify services to 4 app services + the DB.
- Merge ingest_movement + ingest_events into a single ingest_worker:
split each poller's main() into reusable startup_catchup()/register_jobs()
and drive both from one schedule loop in new ingest_worker_rev.py
(standalone entrypoints retained for local debug).
- docker-compose.yaml: replace the two poller services with ingest_worker;
remove the pgbouncer service (dormant; transaction-mode pooling is unsafe
for the advisory-lock'd v_trips refresher) and the grafana service +
grafana-data volume (redundant with the FleetOps SPA).
- Add reporting.v_ingest_health (migration 19) + dashboard_api GET
/health/ingest as the pipeline-freshness surface that replaces Grafana's
health panels.
webhook_receiver stays isolated so a poller fault can't drop inbound pushes.
timescale_db and db_backup are unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
76f6915e61