Join Prathamesh Sonpatki on LinkedIn Live as he revisits the “pillars of observability” — and argues the framing we picked up in 2017 is overdue for an update.
What we’ll cover
- It’s not Logs vs. Metrics vs. Traces — it’s logs and metrics and traces and events, and more. Why each one carries different information and why the “pick one” framing is wrong.
- Why we do observability in the first place — prevent downtime, communicate impact, debug fast, recover, do RCA. The four real stages of an incident, and which signal helps with each.
- The four signals, signal by signal
- Logs — descriptive, easy to adopt, hard to standardize, explode in volume
- Metrics — aggregated, dimensional, best for patterns and anomalies
- Traces — directional, cross-component, OpenTelemetry is making them tractable
- Events — structured, dimensional, the missing piece for correlation
- What it takes to actually run this in production — cardinality limits, flow control, streaming aggregation, data tiering, and the rest of the operational story behind a real telemetry platform.
Bring your questions. We’ll wrap with live Q&A.
