- What happens when a dependency starts failing? What if it begins failing slowly?
- How can the system degrade in a graceful manner?
- How does the system react to overload? Is it “well conditioned?”
- What’s the worst-case scenario for total failure?
- How quickly can the system recover?
- Is delayable work delayed?
- How do you monitor the system? How do you detect anomalies?
- How do you deploy the system? How do you deploy in an emergency?
- Are you learning from all failures?
I hereby claim:
- I am dmathieu on github.
- I am dmathieu (https://keybase.io/dmathieu) on keybase.
- I have a public key whose fingerprint is 18F1 22A4 AE59 73DB FE1A E770 C638 5398 CF53 D518
To claim this, I am signing this object: