Problem Definition

Stability and up-time are critical aspects in production environments for Enterprise clients. Often, we see one of the features from an add-on consume a large amount of resources causing Confluence to crash (OOME) or cause high CPU and thus Confluence becomes unavailable.

Suggested Solution

Implement a configuration option to set per-plugin (per-addon) resource limits (and monitoring) to prevent the entire instance from becoming unstable or unavailable due to insufficient resources.

Workaround

On these types of cases the process typically includes:

1. Reviewing the support zip and thread dumps
2. Analyze a heap dump
3. Identify the problem
4. Disable/troubleshoot the add-on causing the issue