We have started seen un-usual behavior on data base recently where it is eating up complete memory assigned to server. we are no doubt inserting huge amount of data but not getting why it has started now only. we have change storage to fastest storage we got but it is still going behind memory. The pick starts at certain time and ends after 3 to 4 hours. is there any way we can make influx to use the disk more than memory.
Hello @OnkarUmarani
Welcome! Unfortunately this isn’t an uncommon issue for v2 users. Things like cache snapshots, compaction processes, and query load can contribute to this. So if you have flux tasks that are all wrtiing data at the same interval this could contribute to this or complicated flux queries.
You can try messing with these settings:
--storage-cache-snapshot-memory-size=26214400
--storage-cache-max-memory-size=1073741824
--query-max-memory-bytes=104857600
As well as monitoring your series cardinality which could cause memory issues.
All that being said v2 is unfortunately known for this. You might consider upgrading to InfluxDB v3 Enterprise which has better memory management features, particularly for handling historical data.
