2023-02-01: Some prometheus metrics are missing
Current Status
This is an internal observability incident. No direct customer impact, but the reduced observability impacted our ability to troubleshoot.
Scope: The only affected metrics were the rules-based metrics that thanos rule
computes from lower level metrics. This includes apdex, among other things. All other metrics were unaffected and have uninterrupted data.
Cause: The problem was our thanos rule
processes abruptly increased their memory demand and saturated their available memory capacity.
Status: Growing the VMs (and implicitly their memory cgroups via their systemd unit) allowed the thanos rule
processes to resume their work. We also reverted the most recent MR to thanos rule definitions, which may potentially have been related to the increased memory demand. Now that memory capacity is increased, that MR may be safe to re-apply (as it may have just been the last straw rather than a large demand increase). We will reassess that separately from this incident.
📝 Summary for CMOC notice / Exec summary:
- Customer Impact: None
- Service Impact: ServicePrometheus ServiceThanos - Observability metrics in general are missing for the last couple hours.
- Impact Duration: 2023-02-01 15:30 - 17:20 UTC (2 hours)
- Root cause: RootCauseSaturation
📚 References and helpful links
Recent Events (available internally only):
- Feature Flag Log - Chatops to toggle Feature Flags Documentation
- Infrastructure Configurations
- GCP Events (e.g. host failure)
Deployment Guidance
- Deployments Log | Gitlab.com Latest Updates
- Reach out to Release Managers for S1/S2 incidents to discuss Rollbacks and/or Hot Patching | Rollback Runbook | Hot Patch Runbook
Use the following links to create related issues to this incident if additional work needs to be completed after it is resolved:
- Corrective action ❙ Infradev
- Incident Review ❙ Infra investigation followup
- Confidential Support contact ❙ QA investigation
Note: In some cases we need to redact information from public view. We only do this in a limited number of documented cases. This might include the summary, timeline or any other bits of information, laid out in out handbook page. Any of this confidential data will be in a linked issue, only visible internally. By default, all information we can share, will be public, in accordance to our transparency value.