(1/2) Something really cool we do for our new services:

We have a simple but powerful observability framework that makes it really easy to 1) say what you want to instrument and measure, and to 2) make sure you're calling it correctly. https://i.imgur.com/NAgb7vs.png

Building https://center.app. 

https://omarish.com

There's nothing worse than realizing you called it the wrong thing and now your statsd has event-name and EVENT_NAME. With this, event types and timers are strongly typed and defined. https://i.imgur.com/Q7YVD7w.png

Metrics and observability (o11y) are actually pretty easy to set up but so easy to break, sometimes silently. 

And 2 years in nobody is comfortable touching the "critical_path.timer" that someone wrote a long time ago, so metrics and data just build and become less valuable.