Behavior while scaling cluster

[abergs]

Hi,

I'm looking for information regarding cluster behavior while scaling. What I've seen in my tests is that query performance is severely impacted (I'd call it downtime) for a couple of minutes when initiating both a scale-out and scale-in.

Is there any official information on the behavior while scaling?

[#2 · Accepted solution]

@abergs 

Hi,

During scaling operations that cluster goes through deployment; therefore it is not answering queries for a short time and does not ingest data. The data queued for ingestion will be ingested as soon as the cluster is up again.

I'll add this to one of our official docs.

Thanks,

Gabi

[#3]

Follow up question: Is the ingestion endpoint affected by this re-deployment (with failures) or is the ingestion simply queued and delayed?

[#4]

@abergs 

The ingestion is delayed until the cluster is up and running and then it continues from the same point.

Thanks,

Gabi