Common Mistakes and SLO Best Practices

Observability: the present and future, with Charity Majors

Topic 8 of 14

Published Jan 22, 2025 · Clip 43:38–48:19 (4:41)

Charity lists common mistakes: waiting until production breaks to adopt observability, and over-relying on static dashboards that don't let you ask questions. She champions using SLOs as the entry point for debugging—'SLOs are the API for engineering teams'—citing Honeycomb's practice of killing a Kafka node every Monday using SLO budget to test recovery. SLOs also hedge against micromanagement and enable negotiation, and they should be derived from the same data used to debug.