The 37-minute Tuesday: anatomy of a cascading failure
A retry storm, a shared connection pool, and the graph everyone screenshotted. What actually happened.
A retry storm, a shared connection pool, and the graph everyone screenshotted. What actually happened.
The B-tree analogy everyone learns conceals the three behaviours that actually bite in production.
Distributed systems agree on almost nothing — except, mostly, the time. Here's the 40-year-old reason.
If you can't deploy one service without a meeting, you have re-invented the monolith over HTTP.
'There are only two hard things…' — the punchline hides a real taxonomy worth learning.
The originals are shorter, clearer and funnier than the blog posts about them.