When a deploy fails

A failed deploy means your latest change did not go live. Your application keeps running on the previous release — nothing was taken down — so visitors are still served the last version that deployed successfully. Only the new release was held back.

What to do

1. Open the operation and read the summary. The failed deploy records where it stopped — building dependencies, running migrations, or starting the application. That tells you which part of the change to look at.
2. Check your most recent commits. Most failed deploys come from a change in the code itself: a dependency that won't install, a migration that won't run, or a build step that errors. Reproduce the failing step locally if you can.
3. Confirm your environment settings. A deploy that worked before and fails now often points at a changed setting — a new required value, a service that isn't connected, or a credential that expired.
4. Deploy again once you've made a fix. strackt will retry the whole process from a clean state. The previous release stays live until the new one succeeds.

If deploys keep failing after a clean change, reach out and we'll help you read the operation in detail.