We created a Flink application which emits data at an interval (using Timers) of every 15 minutes and store some events in cache for business logic processing these events also gets cleared up after approx 8-10 hours, but we are observing a linear growth in savepoint size, below are the configurations
- checkpointInterval : 7m
- savepointInterval: 12m
- savepointHistory: 96 //approx 1 day of checkpoint history stored
Can we inspect the savepoint data? and what should be contributing in this huge cache size?
Below is the linear increase in Flink savepoints size: