---
title: "Find the deploy a CI deploy step could not match | Proofline Help"
description: "A proofline ci deploy step exits 2 when Proofline recorded no deploy of the commit, or recorded deploys of it and none matches the step's filters."
url: https://proofline.sh/help/ci-deploy-no-matching-deploy/
---

[Help center](/help/)

# Find the deploy a CI deploy step could not match

A `proofline ci deploy` step exits 2 when Proofline recorded no deploy of the commit, or recorded deploys of it and none matches the step's filters. The last line of the step's log says which. Either way the result is not established: it says nothing about how the deploy went, and the pipeline decides what the job does next.

## No deploy was recorded

The log ends with "No deploy of REPOSITORY at SHA was recorded after N minutes."

Just after a deploy starts, no record is the usual answer. When the step ran beside a deploy that was still rolling out, a longer `--timeout` may be all it needs.

Otherwise, find the deploy Proofline recorded:

1.  Open **Environments** and select the environment.
2.  Select **Deploys** in the environment sidebar.
3.  Open the deploy the job made.
4.  Compare the repository the page names, and the full SHA of the commit it links, with the repository and SHA in the log.

A step in another workflow, or one that sets its commit, can read a commit the deploy did not run. Pass the commit the deploy job deployed, as a full 40-character SHA.

When the tab lists no deploy of the commit, check what Proofline can record:

-   The repository is [linked to the environment](/help/repository-environments/) with the GitHub Environment the job deploys to. The tab lists only deployments recorded after the repository was linked, and a workflow-only link records no deploy.
-   The deploy ran through a GitHub deployment. A GitHub Actions job that names an `environment` creates one. A deploy script or tool must create its own and [report its success](/help/aws-github-deployment-status/). The step reads GitHub deploys only, so an AWS CodePipeline execution or Amplify job is never listed, even when the Deploys tab shows it.

## Deploys were recorded and none matches

The log ends with "Proofline recorded N deploys of REPOSITORY at SHA, and none matches --environment, --task, --run or --attempt after M minutes."

Proofline holds deploys of this commit, but at least one of the step's filters differs from each of them. Each filter must equal the deploy's value exactly, including case. Open each deploy of the commit on the **Deploys** tab. Below its status, the page names the repository, then the GitHub Environment, the deployment task, the workflow and job, and the attempt. Compare:

-   `--environment` with the GitHub Environment. This is the name GitHub uses, not the Proofline environment's name.
-   `--task` with the deployment task. The page leaves out GitHub's default task, `deploy`.
-   `--run` with the workflow run. Under **Jobs**, each job links to its run attempt on GitHub, and the run ID is the number after `/actions/runs/` in that link.
-   `--attempt` with the attempt. The page names one only after a re-run, and **Jobs** names it beside each re-run job.

Running the step with `--json` prints every recorded deploy of the commit with these four values, which also gives the run ID of a deploy of one job. [The CI reference](/docs/ci/) describes the options.

### Read from a separate workflow

In a reader triggered by `workflow_run`, use the completed deploying run's `github.event.workflow_run.id` and `github.event.workflow_run.run_attempt`. In a manually dispatched reader, pass the deploying run's ID and attempt as workflow inputs. [See both workflow examples](/help/ci-deploy-later-check/).

### Match a retried job

A job that failed and then succeeded on a later attempt joins the deploy its run's other jobs landed, so that deploy matches the retry's attempt as well as the first. A re-run of a job that had already landed, such as a rollback that re-runs a whole run, is a deploy of its own with its own page and checks, and matches only its own attempt. When the step's attempt is newer than every deploy listed, its deploy may not be recorded yet, and a longer `--timeout` lets the step wait for it.

## Keep the filters

Do not remove a filter so that the step finds a deploy. Another run's or another task's deploy of the same commit is not this job's result, and its checks say nothing about what this job deployed. Correct the value that differs from the deploy page, or give the job's own deploy time to be recorded.

## Related guides

-   [Read a deploy's verification](/help/read-deploy-verification/)
-   [Trace a deployment back to its GitHub Actions jobs](/help/deploy-grouped-workflow-jobs/)
-   [Read Proofline results in CI](/help/read-proofline-results-in-ci/)
