About this event
3 common OTel problems (and how to fix them)
OpenTelemetry isn’t a switch you flip. It’s a tightrope you walk. If you don’t manage the complexity, you drown in performance issues and costs. Keep your set up too simple, and you’re blind during incidents. In this live session, Annie Freeman (DevRel, ex-platform engineer) breaks down how to keep your OpenTelemetry collector set up useful and sustainable, so you can debug end-to-end, keep budgets predictable, and extract more real business value from your telemetry data.
We’ll get concrete on the three things that usually knock teams off balance: sampling (head vs tail plus configuration tradeoffs), cardinality (why queries get slow and bills spike, plus how to stop it), and end-to-end tracing (how to avoid broken traces in Kubernetes/Lambda, including load-balancing and span links for async).