Hi, sometimes our FE instances go into some sort o...
# questions-and-troubleshootinga
Ariel Moraes
11/12/2025, 2:16 AMHi, sometimes our FE instances go into some sort of loop making the BEs use all the CPU available, we checked for loads, queries and connections not in sleep. We couldn’t find anything. The only thing that made the problem go away was restarting the FE nodes. What can be causing that? We are on 4.0
k
Kevin Cai
11/12/2025, 3:49 AMcheck the BE cpu profiling, the query pool mem tracking, what kind of workload running on the BE.
a
Ariel Moraes
11/12/2025, 10:43 AM@Kevin Cai I ran the command SHOW USAGE for the resource groups and it showed some running queries, but when I went to the current queries, nothing was running,
As for the mem tracking I’ll do that if the problem comes to hunt me again. For now the restart of FE nodes made the problem go away.
k
Kevin Cai
11/12/2025, 9:32 PMmy best guess is the backgroup table statistics collection job. check if any table has a high compaction score of its partitions, which might cause the stats collection cpu intensive.