Event Data Evidence Log
Legacy
Area | reporting-monitoring |
Product Manager / Owner | Bryan Vickery |
Tech Lead | Joe Wass |
Language | JSON |
Description | Log of actions taken by Agents, Percolator, Quality. Stream of data, snapshotted. |
Documentation | [https://www.eventdata.crossref.org/guide/service/evidence-logs/] |
Quality | No Sentry, no SONAR |
Upstream services | |
Upstream data | |
Downstream services | |
Downstream data | |
Source Code | |
Products |
Evidence Logs are produced by Agents and Percolator. They are written to a Kafka topic. These logs are structured and intended for public consumption. For more information see the User Guide. The are different to Application logs because:
- Every entry is structured JSON with defined fields.
- Every log type is documented in the User Guide.
- They contain only public log messages that will be of interest to users trying to understand the data.
These are made available to the public via hourly snapshots.
Historically this feature was known as ‘status’, so you may see references to this in the source code or configuration parameters.
Monitoring
The Evidence Log plays two roles in monitoring, both via Heartbeats:
- The logs are monitored by heartbeats. The Evidence Log shows live activity, so, for example, if an Agent stops working this will be detected.
- The logs are snapshotted on an hourly (previously daily) basis See: Periodic Tasks. These are checked by heartbeats URL checks. If a snapshot is missed, it will generate an Error via the Quality heartbeat.