Pipeline freshness
Per-feed observability, updated continuously. The backend recomputes a freshness
snapshot every minute; this page re-reads it on a short interval. Status is computed
from lag against thresholds defined in
ingestion/assets/ops.py.
Loading...
How to read this
- healthy
- — lag within expected polling window
- lagging
- — slow but flowing; investigate if persistent
- stale
- — significantly behind; likely upstream issue
- down
- — no recent records; ingestion broken or feed offline
If you're seeing all "down" — that's the system being honest. Don't trust dashboards that are always green.