Every second, more than 300,000 requests pour logs, metrics and traces into systems that I build.
I'm Tze Yiing, Ziinc on GitHub. At Supabase I lead the observability systems and Logflare development, which takes in over 25 billion requests a day.
Most writing on distributed systems stops at the diagram. Mine starts where the diagram breaks: the global name server stuck in a bootloop, the message queue that will not drain, the garbage collector quietly eating a node's memory. Every post comes from a real incident, with the code and the pull request that fixed it.
If you run Elixir or Erlang in production, you will meet these problems too. Better to meet them here first.
My creations:
- BencheeAsync - Async benchmarking library for Elixir.
- memlite - SQLite-based state management library in TypeScript.
- Crawldis - Declarative web crawler.
Stuff I've contributed heavily to:
- Logflare (Supabase Analytics), lead dev.
- Supabase Studio. Logs explorer, analytics server, log drains.
- Crawly - Web scraping library for Elixir, ex-maintainer, guest featured on the Thinking Elixir podcast.
I take on part-time consulting work. Email [email protected].