Observability
What is it?
Observability is the ability to understand what's happening inside a system from the outside, using its logs, metrics, and traces.
Explain like I'm 5
Why was it created?
As systems grew complex and distributed, simply checking 'is it up?' wasn't enough. Observability emerged to answer why a system behaves as it does.
Where is it used?
- Production monitoring
- Debugging incidents
- Performance analysis
- Understanding distributed systems
Why should developers care?
When production breaks, observability is how you find out why. It's a core skill for running modern services.
How does it work?
Systems emit three main signals: logs (event records), metrics (numeric measurements over time), and traces (the path of a request across services). Tools collect and correlate these so you can ask new questions about behavior.
Real-world example
When checkout slows down, engineers use traces to see which service is slow, metrics to confirm the spike, and logs to find the specific error.
Common use cases
- Incident investigation
- Spotting performance regressions
- Tracing requests across services
- Alerting on problems
Advantages
- Faster root-cause analysis
- Understand complex systems
- Catch issues before users do
- Answer unforeseen questions
Disadvantages
- Data volume and storage costs
- Instrumentation effort
- Too many alerts cause fatigue
When should you use it?
For any production system, especially distributed ones, where you must diagnose issues quickly.
When should you avoid it?
Rarely — even small systems benefit; extremely simple scripts may not need full tooling.
Alternatives
Related terms
Interview questions
Beginner
- What are the three pillars of observability?
- How is it different from monitoring?
Intermediate
- What is a distributed trace?
- What's the difference between logs and metrics?
Senior
- How do you reduce alert fatigue while keeping coverage?
- How would you instrument a service for observability?
Common misconceptions
- "Observability is the same as monitoring" — monitoring watches known signals; observability lets you ask new questions about unknown problems.
- "More dashboards means more observability" — quality and correlation matter more than dashboard count.
Fun facts
- The three pillars are logs, metrics, and traces.
- The term comes from control theory, about inferring internal state from outputs.
Timeline
- 2010s — Observability gains traction with microservices and cloud
Learning resources
Quick summary
Observability is understanding a system's internal behavior from its logs, metrics, and traces, so you can diagnose problems and answer new questions.
Cheat sheet
- Three pillars: logs, metrics, traces
- Understand behavior from outside
- More than uptime monitoring
- Key for debugging distributed systems