Shutdown of digest had a timeout of 10 seconds internally, which could lead to being dropped too soon while shutting down or restarting. This could result in ingest lag rising to over .0 seconds, where the expected lag is the time from the shutdown is initiated until a fe seconds after the new instance is started. There is a new config SHUTDOWN_ABORT_FLUSH_TIMEOUT_MILLIS which defaults to .0.0 millis (5 minutes) to allow proper shutdown also on systems with many datasources or slow filesystems / disks.