Now that swh.scheduler tries to send statistics to a statsd instance, we should probably make sure that one is available somewhere. We also need prometheus to pull the data from that instance.
Description
Description
| Status | Assigned | Task | ||
|---|---|---|---|---|
| Migrated | gitlab-migration | T1408 More/better Metrics | ||
| Migrated | gitlab-migration | T1535 Deploy prometheus-statsd-exporter to gather per-worker metrics | ||
| Migrated | gitlab-migration | T1460 Add task related metrics to swh-scheduler | ||
| Migrated | gitlab-migration | T1461 Add loader-related metrics to swh-loader-core |
Event Timeline
Comment Actions
statsd-exporter now has a puppet manifest to be deployed to all machines.
rSPSITEd6dd9660f20f73b03185bc6c4dece0c5eaeb0216
Deployment on all workers has been done. As the new base code for tasks has been deployed a while ago, it's already getting some data.