Prometheus has tsdb config for block duration (
min-block-duration ) which dump/flush the head to TSDB block on disk on regular interval (2h by default).
I want to know is there a way (api/cli) through which we can force this flush operation on demand instead of waiting for 2 hours all the time.
Motivation: I am planning to use Prometheus with local PV and thanos with objectstore copy. During a node rotation, I want to force flush in-memory data (which is not flushed to TSDB block yet due to 2h interval) in Prometheus to objectstore (thanos side car copies TSDB block to objecstore) such that no data is lost when the promethues pod moves from one node to other node with empty local volume.