Frequent GetObject Requests to S3 in Elasticsearch Snapshot Repository

Elastic Stack Elasticsearch
1

Hello,
I’m encountering an issue with GET Object operations while using Elasticsearch snapshots, in aws s3 bucker repository
and I’d appreciate some guidance.

Currently, I’m trying to snapshot an index that has a size of around 600 MB. I noticed that there are approximately 600 GET Object operations occurring. Surprisingly, even when I’m not performing any snapshot or restore operations, I observe 20-30 GET Object operations happening every 5 minutes.

Previously, I had around 40 snapshots with indices ranging between 300-400 MB in size. During that time, my GET Object requests were much higher—around 4000-8000 operations within 30 minutes.

I’m puzzled as to why there are so many GET Object operations, even when I’m not actively snapshotting or restoring anything. Could someone help explain why this might be happening and how I can address or optimize it?

Any insights or recommendations would be greatly appreciated!

e472b049-17d0-4693-8d1b-cbc046c752eb
e472b049-17d0-4693-8d1b-cbc046c752eb1694×901 238 KB

Thank you!

2

Elasticsearch only retrieves data from the repository when specifically told to, so I would guess that something is hitting the get snapshot or get snapshot status APIs. Maybe a rogue monitoring system?

Taking a snapshot requires reading just one blob per shard IIRC so that wouldn't explain it.