Eight definitions of “observability” vs. “monitoring” – and why none of them work
summary
Eight proposed distinctions get evaluated with a simple CUP rating (correctness, usefulness, popularity) and none pass without flaws: the control theory definition is mathematically correct but irrelevant to most ops, the "what vs why" split collapses into alerting, the three pillars (logs metrics traces) is catchy but incomplete, and vendors weaponize Prometheus monitoring vs observability. Conclusion: observability works best as an exploratory property or outcome of monitoring, not a replacement, and Prometheus remains a valid monitoring tool within an observability toolkit.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 3:53Observability buzzword rise
- 10:00Control theory definition
- 13:51Monitoring vs observability
- 17:44Distributed tracing confusion
- 18:48Three pillars myth
- 21:17Hipster monitoring
- 24:05Marketing weapon
- 25:52Meaningless buzzword
- 27:39New observability definition
- 31:56Why definitions fail
- 34:28Q&A
related talks
35mObservability: Building Confidence & Answering the What, Where, How
40mDemystifying Observability and Monitoring
06mIgnite | Observability vs Monitoring - A Linguistic Analysis
05mObservability will not fix your broken monitoring
29mObservability; a gentle introduction
39mThe 5W’s of Observability with Practical Examples
30mObservability is More than Logs, Metrics & Traces
29mPanel - Observability: What's the big idea?