> For the complete documentation index, see [llms.txt](https://developer.harness.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://developer.harness.io/continuous-delivery/troubleshooting-and-resources/armory/general/ecs-deployment-errors-with-all-server-group-names-for-cluster-are-taken.md).

# ECS Deployment errors with "All server group names for cluster....are taken"

### Issue <a href="#issue" id="issue"></a>

A transient issue occurs with the ECS deployment stage, across multiple pipelines within the same application. This issue can occur even across different server group names. The following error is observed: `All server group names for cluster .....are taken.` `All server group names for cluster ...are taken.` The UI shows the following:

### Cause <a href="#cause" id="cause"></a>

The issue is as a result of hitting a limit on the number of possible server group names in Spinnaker (e.g. there may be over 1000 versions of a specific serverGroup), and the `NameBuilder` that is capped at `1000`. <https://github.com/spinnaker/clouddriver/blob/01f415f318a15970729b4415167e64eafeca9593/clouddriver-core/src/main/groovy/com/netflix/spinnaker/clouddriver/helpers/AbstractServerGroupNameResolver.groovy#L31> It is very rare to hit this limit of 1000 ECS server groups within a cluster, and it does not impact every v2.24 environment.