I have a warning on Monitoring console of my cluster master, that my Index disk utilisation is critically high. I have a 30 day retention policy set on all my indexes however the disk utilisation is continuing to rise.
In monitoring console - Indexing - indexes and volumes - instance - Indexes table.
I am seeing Data Age vs Frozen age that I have data much older than my retention policy.
To me that would suggest I have hot buckets that have not rolled over perhaps due to data ingest latency or incorrect timestamps preventing the roll over.
Could this be the source of my increasing disk utilisation?
My understanding is that if a rolling restart of my Indexers occurred then all buckets would roll over, however I know there has been a restart within the time frame indicated by the data age.