Hello, I detected a problem with my flink applicat...
# troubleshooting
f
Hello, I detected a problem with my flink application where the memory is growing in the job manager constantly, it looks like a memory leak, it is implemented with apache flink 1.17.2 , the program is so simple because consumes data from rabbitmq source and stores in starrocks sink, do you know if there is a known troubleshooting for this? I am using flink kubernetes operator and there is no additional configs, it is enabled with checkpointing each 5s, at least once and with ha enabled storing in a s3 bucket
p
Hello there @Fabrizzio Chavez What is the size of your checkpoints?
f
Hello @Pedro Mázala, is less than 3kb, the checkpointing is only used to confirm the ack needed by rabbitmq and the commit write for starrocks sink
p
That is weird. Was it always like that? Or something changed?