Observability in practice, so that issues have nowhere to hide

Hook

“I can’t reproduce it,” “we’ll need to add logs,” “we can’t test that in staging,” “it seems to have been broken for a while,” “it works locally,” …
If you hear these phrases regularly, this talk is for you.
If you often feel like you’re a step behind on bugs, this talk is also for you.

Summary

During this talk, I introduce the concept of observability and how it can help developers always stay ahead, increase application quality, and thus have more time to add value!
I then present the concrete methods and tools we have implemented at WP Media to maintain and monitor our products, which are used on over 4 million websites, and our services, with daily production deployments and several thousand requests per second; all managed by only a dozen developers.

I briefly revisit the basics, but enhanced:

  • Standardized logs that anyone can add,
  • On-demand test environments

And I discuss more advanced methods and tools to simplify their implementation:

  • Error & Performance tracking, Tracing, and technical alerts with Sentry
  • Functional alerts and data observability with Metabase
  • Canary releases
  • Shadow releases
  • Health Probes

If these techniques are unfamiliar to you, if you have difficulty implementing them, or if you simply want a refresher, this talk is a great opportunity to hear about practical experiences and tips on observability!

Main takeaways

  • Monitoring vs. Observability
  • Standard log implementation for everyone
  • Error reporting, Tracing, Telemetry
  • Technical vs. Functional Alerting
  • Tooling: Metabase, Sentry

Targeted audience

This talk is primarily designed for developers, team leaders, and engineering managers. More senior tech profiles can also benefit from this talk. Attendees with previous experience with back-end services will fully benefit from this talk.