Ceph Crash Telemetry - Observability in Action
summary
Ceph ships a built‑in Telemetry module that 2500 clusters currently opt into, reporting 2500 clusters and nearly 1 exabyte and about 130k OSDs; collection uses five channels (basic, crash Channel, device, ident, perf) with drive SMART and performance counters; back end stores anonymized UUIDs (no fsid or IPs) and groups raw backtraces by regex normalization, then a Telemetry bot opens or updates Redmine issues so fixes and regressions are tracked and verified via public Grafana dashboards.
AI-written summary · about
chapters (15)
- 0:00Introduction to Ceph Telemetry
- 1:29How Telemetry works
- 5:04Why Telemetry matters
- 8:19Telemetry architecture
- 9:01Public dashboards overview
- 12:17Drive health metrics
- 13:44Crash data processing
- 16:37Crash signature grouping
- 19:06Crash dashboard demo
- 22:41Crash triage and QA
- 27:41Community-driven fixes
- 29:10Feature adoption insights
- 30:17Opting into Telemetry
- 30:40Future improvements
- 34:58Q&A session
related talks
38mEnhancing Ceph Monitoring at CERN: Our Approach and Solutions
27mInsightful Traffic Monitoring: Harnessing Cilium for Comprehensive Network Observability
22mReworking Observability In Ceph
34mDay 2 Operations : Make Friends with Your Ceph Cluster
33mMonitoring Ceph with Prometheus
30mPanel: Telemetry That Matters
32mEnhancing Observability and Monitoring for Large Ceph Clusters at Scale
38mEfficient Ceph Performance Troubleshooting in Production Using eBPF