Missing index after migration: pre-release and first 30 minutes checklists

A migration drops or forgets an index and query latency rises under production load.

The pre-release checklist maps each item to a test class from the incident page. The first-30-minutes checklist covers detection, containment, the rollback decision and customer communication.

8 pre-release items · 7 first-30-minutes items · verified 2026-10-21

Pre-release checklist

PR-01List every DROP INDEX in the migration and confirm a valid replacement index exists before the drop runs. migration
PR-02After each CREATE INDEX CONCURRENTLY step, query the index validity flag and stop the migration if the index is INVALID. migration
PR-03Compare the index list on each changed table before and after the migration on a production-shaped copy, including indexes on a replaced primary key. migration
PR-04Ship the drop of a replaced index in a later release than the creation of its replacement. migration
PR-05Run the query plan for the most frequent queries on the changed table against the migrated copy and confirm each one still uses an index. migration
PR-06Replay the high traffic queries against the migrated copy under load and compare P50 and P75 endpoint latency with the pre-migration baseline. load
PR-07Apply the migration to one shard or region first and hold promotion until database CPU and query latency stay at baseline. canary
PR-08Hold promotion while background job latency or primary read rate on the canary shard exceeds baseline. canary

First 30 minutes

FM-01Compare the index list on the slow table with the list from before the migration.
FM-02Check whether any index on that table is marked INVALID.
FM-03Stop the progressive rollout of the migration to the remaining regions or shards.
FM-04Pause analytics queries and other non-essential workloads on the table before rebuilding the index.
FM-05Decide whether to rebuild the index now or after peak traffic, because a rebuild under load can overload the database cluster.
FM-06Recreate the index shard by shard, and on shards under heavy load revoke application reads and writes on the table during the build.
FM-07Publish a status note listing the slow or failing features, such as API timeouts or delayed background jobs.

Download the checklist

Pre-release checklist (Markdown)1.6 KB · v1.0.0
First 30 minutes checklist (Markdown)1.1 KB · v1.0.0

Only in this checklist

Every test class in this checklist also appears in at least one sibling checklist in the group; the item texts differ: List every DROP INDEX in the migration and confirm a valid replacement index exists before the drop runs.

Questions

What should be checked before a release to prevent Missing index after migration?

8 pre-release items, PR-01 to PR-08. PR-01: List every DROP INDEX in the migration and confirm a valid replacement index exists before the drop runs.

Is there a rollback plan for Missing index after migration?

The first-30-minutes checklist has 7 items. FM-01: Compare the index list on the slow table with the list from before the migration.

What should a team do in the first 30 minutes after Missing index after migration is detected?

The first-30-minutes checklist has 7 items. FM-01: Compare the index list on the slow table with the list from before the migration.

Is a Missing index after migration checklist still needed if deployment is automated?

Yes. Automation runs the steps; the checklist names what the tests must cover: migration tests, load and performance tests and canary and progressive delivery checks.

Which pipeline stages does the Missing index after migration checklist cover?

The pre-release items map to the pre-deploy and post-deploy stages.

In what formats can the Missing index after migration checklist be downloaded?

Both checklists download as Markdown files: /downloads/missing-index-after-migration-pre-release.md and /downloads/missing-index-after-migration-first-30-minutes.md, licensed CC BY 4.0.