Skip to content

Deployments are records, so rolling back is a choice in a list

Each deployment keeps its commit, its build log and its outcome. Going back means picking an earlier row, not rebuilding from memory.

4 min readREM CloudHub team

DeploymentsOperations

A deploy that goes wrong at the end of the day is stressful mostly because of the questions around it. What changed? Which commit is live? What was running before? REM CloudHub answers those before anyone asks, by keeping every deployment as a record.

What a deployment keeps

  • the project and the branch it came from
  • the exact commit
  • the build log, line by line
  • who started it, and how: from the console, the API or an agent through MCP
  • how it ended: running, failed during the build, or failed to start

A failure says why

When a build fails, the deployment says which step failed and shows the last lines of its log. When the container fails to start, Activity records it with a reason, and the error map turns the error code into a plain sentence with a reference you can search for.

Deploy failed. Why: pip install exited 1. Reference: req_7f3c…

Rolling back

Rolling back is the same action as deploying, pointed at an earlier record. Pick the last deployment that was healthy and RCH runs it again at the same address. The bad deployment stays in the list with its log, so the investigation can wait until morning.

Working with it from scripts

Everything on the deployments page is also in the REST API and the MCP server, so a script or an agent can find the last healthy deployment and run it again. The exact routes are in the API reference that ships with your installation.

Was this useful?

Subscribe to the newsletter

New posts and product news from REM CloudHub, a few times a month. Nothing else.