Performance during Snapshot Backups/Cross-Cluster replication

Elastic Stack Elasticsearch
1

Assume the cumulative Index Size is approximately 200GB.

Is taking snapshot backup every 30 minutes advisable? Will it cause any performance impact on the cluster for any active read/writes happening to the cluster during the snapshot backup process?

Similarly, will the cross-cluster replication cause any performance degradation to the primary cluster's read/writes ??

2

When you say:

Assume the cumulative Index Size is approximately 200GB.

Do you mean that the total amount of data in the cluster will be 200GB, or that within the 30-minute window 200GB of data will change requiring a new backup?

Either way, generally speaking, the backup/snapshot process is fairly efficient on newer versions of Elasticsearch, as long as you have a good (fast) network connection to the backup destination, backups generally shouldn't cause many issues related to performance.

3

The total amount of data in the cluster say is 200 GB.
Within the 30-minute timeframe, assume a data of about 10-15 GB is ingested.

When I configure the snapshots to be taken/backed up to remote repository every 30 minutes, how significant would the impact on the performance of read/writes be during the snapshot backup process? I assume it's a separate single thread running that transports the backup, so I believe it shouldn't really affect the usual read/writes.

(Running on a 2vCPU and 8GiG of ram per node on a 3 node cluster.)

4

I unfortunately can't really say. I would not expect much of a performance impact, but the only way to know for sure is to actually test it and see what happens.

5

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.