Open KalmanMeth opened 1 year ago
In the client, any metrics we want to stop reporting must be explicitly deleted. In the old implementation of encode_prom there was a bug causing the data to never be cleaned up. This seems to be fixed in the more recent implementation. https://github.com/netobserv/flowlogs-pipeline/pull/266
By default, Prometheus metrics are stored for N days (based on configuration, typically 7 or 15 [AWS]). Can set retention period and size in config file by setting the storage retention options.
--storage.tsdb.retention.time=30d
--storage.tsdb.retention.size=10GB