Changefeed resolvedTs lag on dispatcher manager node does not decrease after restart #4203
Description
Issue Description
The resolvedTs lag of a changefeed on a specific dispatcher manager node fails to decrease after the changefeed is restarted.
Reproduction Steps
A failure occurs in the downstream Kafka cluster.
The changefeed latency increases significantly.
Perform pause followed by resume operations on the affected changefeed.
Workaround
Restart the dispatcher manager node on which the changefeed resolvedTs lag continues to rise.