10 years of OSMC: Why does my monitoring still look the same?
summary
Ten years of conference attendance yields one concrete claim: monitoring tooling shifts (checks → metrics) but fundamental problems remain unresolved. Legacy Nagios "feature complete" forks still run; modern stacks use Prometheus scraping HTTP endpoints and tacked‑on alerting. Common failures: thousands of unlabeled metrics, meaningless histograms over averages not being collected, dimensional chaos across clouds, and no protocol to tell clients the right units or bucket sizes. Proposes integrating logs/metrics/traces and a two‑way monitoring protocol; demos are conceptual rather than finished products.
AI-written summary · about
chapters (5)
- 0:00Introduction and journey
- 14:09Nagios legacy and decline
- 22:18Modern monitoring pitfalls
- 33:08Future of monitoring vision
- 36:01Q&A session
related talks
39mThe Story of One Monitoring
52mHolistic Monitoring
38mo11y will not fix your broken monitoring
23mObservability in OpenStack Learnings and building blocks from the SIG Monitoring at SCS
39mThe Ticking Timebomb of Observability Expectations
40mDemystifying Observability and Monitoring
22mService Monitoring Manual—2018 Edition
21mImproving Monitoring Systems Interoperability with Prometheus & OpenMetrics