Avoiding performance issues during snapshot deployment
Don’t select the Send ClearCacheEvent to clear Jira's internal caches after deployment in the general settings in CMJ. This way, CMJ will only refresh the caches of objects changed during deployment, not all the caches on the system. Jira Data Center replicates its memory cache across all other nodes using a specially configured Ehcache and RMI. User and group permissions are cached in each node's memory, among other data items. This means that if the Send ClearCacheEvent option is selected, a significant temporary slowdown of the system may be experienced by the users. More details about Jira Data Center scaling can be found in this article.
If the system is under heavy load when deploying changes to the user management, the cache replication mentioned above may cause a temporary slowdown. This shouldn't be as significant if the Send ClearCacheEvent option is left unchecked.
What is the effect of CMJ operations on Jira Data Center cluster performance?
Jira Data Center allows you to run a cluster of multiple Jira nodes providing stable performance under high load. The CMJ operations (snapshot creation, deployment, integrity check, etc.) always run on a single node. This is the node where the user is logged in currently and starts the operation from.
Running CMJ operations on a single node ensures that your Jira performance across all other nodes remains stable and no slowdowns in response time occur. Your Jira performance will be affected only on the node where the CMJ operation is running.