I have a flink Streaming job that takes in stream of events as input from kafka. This stream is then filtered, and the events are then collected in EventtimeWindow of 1 min. This 1 minute window is resampled into a single event and then this stream of resampled events is collected into a sliding count window (slide = 1) and the count window is then sinked to Kafka.

Everything is working as expected, except, my taskmanager is getting filled up due to checkpoints not getting cleaned up.

I am triggering checkpoints every minute and i have a num-retained value of 2 in the flink config. I see that job manager is very well respecting this config as i always see the last 2 chk-x files in job manager storage but in task manager storage i see all the chk-x starting from chk-1 till now. And no file is cleaned.

Since i am relatively new to flink and i have been struggling with this issue for sometime now, it would be a great help if someone can point me in the direction of what is happening here.

Thank you!