Customer is reporting long running cleanup task using an S3 blobstore.
In the cleanup task log there are several occurrences of these messages:
2020-07-20 12:57:16,733+0000 WARN [quartz-10-thread-18] *SYSTEM org.sonatype.nexus.cleanup.internal.service.CleanupServiceImpl - Search scroll timed out, continuing with new scrollId.
Each of these warnings is accompanied by a delay of 5-7 minutes. Combined this results in quite a bit of overhead.