Automation of User and Service Account Removal in Continuous Delivery

I would like to propose a significant operational improvement for IBM Cloud Continuous Delivery: automating the process of removing identities, including both users and service accounts associated with GitLab.

Currently, this process is manual and prone to errors, resulting in unintended reactivation of accounts that should have been permanently removed. Besides operational and governance costs, these errors directly impact billing, as the service is charged based on the number of authorized users per instance per month.

Automation would bring consistency, security, efficiency, and cost optimization—preventing unnecessary charges for users who should no longer have access.

Problem Statement

We face recurring cases where users removed from Continuous Delivery are automatically re-added. Joint analysis with IBM Cloud support and TAM revealed this issue is caused by incomplete removal, which must simultaneously address three areas:

• Remove the user from IAM across all toolchains in the resource group.

• Remove the user from the Authorized Users list in the Continuous Delivery service instance.

• Revoke Developer access in all linked Git Repos (Git Repos and Issue Tracking) associated with the toolchains

The current manual process is error-prone and susceptible to inconsistencies, rework, and, importantly, billing inaccuracies—as each reactivated user remains counted in the monthly authorized users metric

Proposed Solution

Create an automated workflow (via Console, CLI, API, or toolchain) that:

01) Upon removal of the user from the Continuous Delivery IAM Group, automatically performs:

• IAM revocation across all toolchains.

• Removal from the Authorized Users list in the Continuous Delivery instance.

• Access revocation in linked Git repositories.

02) Sends an alert if any step cannot be completed automatically.

This automation ensures complete, secure user removal and avoids unnecessary charges—while significantly enhancing governance and operational efficiency.