Backup size ratio Cassandra vs. Elasticsearch

[MarcoCritti]

Hi Community!

I was wondering, what are the typical backup ratios for you?

My calculations are based on the reference here: https://www.dynatrace.com/support/help/shortlink/managed-cluster-backup-estimate

Cassandra: (SUM of all Nodes Metric Storage * 0.2) * 2
Elasticsearch: SUM of all Elasticsearch storage
 
OR
calculation based on the first node:
Cassandra: (Count of Nodes * Metric Storage * 0.2) * 2
Elasticsearch: Count of Nodes*  Elasticsearch storage

Is it like 20% Cassandra : 80% Elasticsearch? Or what more like 40% : 60%? Im our case we try to get a feeling for backups and how it will look like in the  future with a fast growing environment, at the same time we want the bare minimum to be in the back up to safe storage which in our case is a cost factor ;).

Thanks to all in advance
Marco

[Mohamed_Hamdy · Accepted solution]

Hi @MarcoCritti ,

as the size might vary slightly between nodes. You can calculate it based on the node with the most consumed storage.

example. Cassandra: (Count of Nodes * Metric Storage of the node with the most consumed storage * 0.2) * 2

Note that this estimate won't be accurate if it was calculated on a new cluster that just started storing data. As the storage on disk grows, so will the size of the backup.