This is reproducible on Data Center: yes
Problem & solution description:
Index should not be overloaded. Actions on issues with large number of comments, worklogs & history should not affect node & cluster performance & stability.
Index is overloaded. When this happens it will affect the performance of all actions triggering re-indexing. Jira may start dropping indexing request which can affect crucial Jira DC components like index replication.
When the index is overloaded you may see the following exception thrown on failed indexing actions:
The problem can be monitored with JIRA-STATS INDEXING-QUEUE logs.
Archive issues with large number of related entities (comments, worklogs, history).