Start from the root cause.
Each incident's root cause is the starting point, so the guardrails target the cause, not the symptom it produced.
- Starts fromThe root cause
- TargetsThe cause, not the symptom
- CoversEvery prod incident
Prevent Recurring Prod Incidents
Fix an incident once. Catch it on every change after. Each production incident generates a regression test and a code review rule aimed at its root cause, so the same failure is caught before it ships again.
From incident to guardrail.
Every prod incident automatically generates a regression test and a code review rule targeting its root cause. This prevents recurring prod incidents.
Incident INC-2041
The incident
Checkout fails for older cartsSEV-1
Checkout returned errors for carts created before
multi-currency support. Their currency field
is empty.
Root cause
A recent change to checkout/totals.py reads
cart.currency without a fallback.
Generated from the root cause
tests/test_totals.pyRegression test
def test_total_for_cart_without_currency():
cart = make_cart(currency=None)
assert cart_total(cart).currency == "USD"Code review ruleTargets the root cause
Code that reads cart.currency must handle
carts with no currency. Carts created before
multi-currency support have none.
A later pull request that reads the field without a fallback is flagged in review, before it merges.
The same incident
should not happen twice.
A postmortem says what went wrong, but nothing stops the same mistake from shipping again. CloudAEye turns each incident's root cause into a test and a review rule, so the fix stays fixed.
A regression test for the failure.
A test that reproduces the conditions behind the incident, so CI fails if the bug comes back.
- TargetsThe root cause
- RunsWith your other tests
- FailsIf the bug returns
A review rule for the pattern.
A plain-language review rule describes the mistake, so every later pull request is checked for the same pattern.
- Written inPlain language
- Applies toEvery later review
- CheckedBefore merge
Guardrails that add up.
Each incident adds a test and a rule. Over time, reviews and CI cover the failures your team has actually seen.
- Per incidentA test and a rule
- Over timeReal failures covered
- ResultFewer repeat incidents
How it works.
It runs automatically for every production incident. There is nothing to trigger.
Read about review rules- A production incident happens.
- Its root cause is identified.
- A regression test and a code review rule are generated.Both target the root cause.
- Later changes are checked.The rule in code review, the test with the rest of your tests.