Collection

Change live systems without turning one mistake into a large one

Make software, configuration and bulk-data changes easier to test, observe, retry, stop and recover before their blast radius becomes the main problem.

15 elements · Explore in any order.

3 checklists · 1 heuristic · 1 pattern · 2 principles · 8 protocols

Narrow your search

15 elements

Protocol

Stage the mutation before committing it

The safest write is often the one you can inspect before it becomes a write.

Read the idea
Protocol

Canary the batch before scaling it

Do not make the millionth write your first realistic test.

Read the idea
Checklist

Write the rollback trigger before deployment

'We will know when to roll back' is not a rollback condition.

Read the idea
Protocol

Test the rollback before you need it

A rollback plan that exists only on paper is a hypothesis.

Read the idea
Pattern

Give a retryable mutation one stable operation ID

A retry should mean 'finish this operation,' not 'please create another one.'

Read the idea
Protocol

Retry slower, with randomness, and stop

A retry storm is an outage trying to help.

Read the idea
Principle

Treat a timeout as an unknown outcome

No response is not the same thing as no effect.

Read the idea
Protocol

Reconcile what changed after a bulk write

'Job finished' is a process status, not a data-quality result.

Read the idea
Checklist

Write invariants before the migration

A migration can be perfectly formatted and completely wrong.

Read the idea
Principle

Shrink the change before shrinking the review

A giant change does not become small because the ticket fits on one screen.

Read the idea
Heuristic

Let the change bake before stacking the next one

If you change the experiment before the result arrives, you inherit an explanation problem.

Read the idea
Checklist

Version the configuration that changes behavior

If configuration can change production, it deserves a history as serious as code.

Read the idea
Protocol

Detect configuration drift instead of assuming sameness

'Same deployment' does not guarantee the same state.

Read the idea
Protocol

Dry-run the real-shaped input without the real write

Real data can teach you something before it is allowed to change anything.

Read the idea
Protocol

Make the rollout stop itself on a known bad signal

A safety metric is more useful when it can pull the brake.

Read the idea