how much storage we can save by enabling frozen bucket concept and roll over data from cold bucket

[ashikuma]

how much storage we can save by enabling frozen bucket and roll over data from cold bucket.

We have tons of data coming in and instead of storing data in cold db for 1 year ,we want to store it in cold db for 3 months max and older than that we will move to frozen bucket , but by enabling that we would like to understand how much storage we can save if we store 9 months data in frozen and retrieve it whenever required.

Example: you can take 3 TB coming in per day ,we have 14 indexers ,multi site cluster (R.F-2, S.F-1).
We would like to keep data till 3 months (max) in hot\warm\cold DB then we would store that frozen DB.
what would be compression factor if we do that , is it gonna save some storage or it will consume same storage that

[richgalloway]

It depends. If you leave the frozen buckets on the same storage device then you're just saving yourself the metadata that is discarded when buckets are frozen. That's about 15% of the original data size.
If you move the frozen buckets to a different storage device then you'll save roughly 75%.
Buckets are already compressed so freezing them does not change the compression factor.