Bow for me for I am Coroot
summary
Coroot claims to be an eBPF-based exporter and a plausible competitor to Datadog/NewRelic, Apache‑2.0 licensed, running on bare metal, Kubernetes or Nomad. It requires a Prometheus‑compatible backend plus ClickHouse for traces and optionally Postgres (or SQLite locally) for project metadata. A node agent auto‑discovers apps and ships eBPF metrics, traces, logs and profiles with almost no manual instrumentation; UI shows service graph, DNS errors, SLO/error‑budget alerts and span‑level traces. Works well in development; production rollout needs project automation and multi‑tenant backend setup.
AI-written summary · about
chapters (17)
- 0:00Introduction and setup
- 1:25Speaker background and events
- 2:52From root to observability
- 4:15Monitoring vs observability
- 6:45Challenges in modern stacks
- 8:31OpenTelemetry and eBPF
- 12:10Introduction to Coroot
- 13:56Coroot architecture and setup
- 17:30Projects and agents
- 19:58Coroot dashboard overview
- 21:44Debugging with Coroot
- 23:07Incidents and SLOs
- 24:14Tracing and logs
- 25:41Alerting and limitations
- 27:04Cultural fit for SLOs
- 28:29Coroot in production
- 29:12Conclusion and call to action
related talks
31mEffortless, standardised homelab observability with eBPF
31mZero instrumentation observability with eBPF
36mBumbleBee: Or How I Learned To Stop Worrying And Love eBPF
19mDiscover seamless observability with eBPF
37mZero-instrumentation observability based on eBPF
33mAuto-instrumentation of Prometheus for RED Monitorin...
27mStep By Step Kubernetes Observability with eBPF
29mRuntime Security at Scale With EBPF: Hybrid Detection and Root Cause…