The best architecture diagram is the one the new hire can redraw from memory. If the error message does not say what to do next, it is half an error message. The rollback plan should exist before the rollout plan is approved.
Going further
Logs answer questions you knew to ask; traces answer the ones you did not. We wrote the migration twice: once to learn, once to ship. Most production incidents start as a small assumption nobody wrote down. Caches are easy to add and surprisingly hard to remove.
Feature flags are debt with an expiration date you must actually enforce. The best architecture diagram is the one the new hire can redraw from memory.
Trade-offs
We wrote the migration twice: once to learn, once to ship. Small pull requests get real reviews, large ones get approvals. Every retry needs a budget, otherwise it is just a slower outage. Most production incidents start as a small assumption nobody wrote down.
- Write the failure mode down before writing code
- Keep migrations reversible
- Alert on symptoms, not causes
- Delete dead code the same week
If the error message does not say what to do next, it is half an error message. Logs answer questions you knew to ask; traces answer the ones you did not.
Getting started
Our on-call rotation improved the day we deleted half of the alerts. Feature flags are debt with an expiration date you must actually enforce. The fastest query is the one you never send, but the second fastest has a good index.
A boring deploy pipeline is a feature, not a lack of ambition.
We wrote the migration twice: once to learn, once to ship. Tests are documentation that fails loudly when it goes stale.
We measured first, and the bottleneck was not where anyone expected. We wrote the migration twice: once to learn, once to ship.
Leave a Reply