Technical craft
Production Ownership
4 lessons · ~25 min total
01
Owning what you ship
~6 minTreat production behaviour as part of the work rather than as something that happens after it.
Key takeaways
- Operational cost stays invisible to an author who never carries the pager
- Closing the loop turns a class of incident into a design change
- Speed of repair matters less than incidents that stop happening
02
Instrument for the question you will ask at 3am
~7 minAdd observability aimed at diagnosis under pressure rather than at dashboards nobody reads.
Key takeaways
- Write the sentence you want to say at 3am, then instrument backwards from it
- Split latency into outbound calls versus own processing
- Aggregates say something is wrong; only dimensions say where
03
Alerts that deserve to wake someone
~6 minAlert on user-visible symptoms with a clear action, and delete the rest.
Key takeaways
- Page only on user-visible symptoms with an action available now
- Alert fatigue is a design defect, not a discipline problem
- Causes without actions belong in graphs and tickets, not pages
04
What to do with an incident afterwards
~6 minTurn an incident into a change in the system rather than a change in how carefully people behave.
Key takeaways
- Ask what made the mistake easy, not who made it
- Action items that ask for more care prevent nothing
- Change what is possible so a tired person cannot repeat the failure
Try a sample check
A sample from the Production Ownership track. The real checks unlock with an account.
Production Ownership
Which of these has earned the right to wake someone at 3am?
Start Production Ownership.
Your first lesson takes five minutes, and you can do it right now without an account.