Today we had grafana-rw become saturated and unresponsive. The VM has been sitting very close to its memory limit for a long time, it needs a bit more breathing room. The allocation should be increased (ideally doubled) given the importance of Grafana.
Description
Description
Related Objects
Related Objects
- Mentioned In
- T414607: Improve Grafana scalability
Event Timeline
Comment Actions
Mentioned in SAL (#wikimedia-operations) [2026-01-14T16:42:44Z] <herron> restarting grafana1002 for memory increase T414604
Comment Actions
Bumped the grafana VMs in codfw and eqiad to 12G (up from 4) and used this as an opportunity to increase VCPUs to 2 as well