Everything is Up! (and other useful myths)
Speaker: Christine Yen
You can't build a great platform without great reliability and great support — and both of those depend on your platform being understandable. Traditional monitoring tells you about the health of the service, but none of your users care about the average request latency or error rate; they care about their OWN request latency and error rate.
Let's talk about where the differences lie, and how the world's greatest platforms use tooling to inform both systems engineering and support. Because your systems are never “up”: complex systems exist in a constant state of partially degraded service.
If your tools don't let you dig in and inspect at the individual request level, they're not going to be good enough in an era where distributed systems are the new normal.
Christine Yen is co-founder and engineer at Honeycomb. She has built systems and products at companies large and small and likes to have her fingers in as many pies as possible.
Previously, she built Parse’s analytics product and wrote software at a few now-defunct startups.