Preview engine runs out of memory

1

We are running the Dremio AWS Marketplace version. At the moment, the preview engine uses 99.96% of allocated memory. As a result, all respective queries fail. At the same time, it shows no CPU activity. I wonder what may have eaten up all that RAM, and how to free it?

An adjacent question: is there a way to configure a separate preview engine node? Currently, it is by default co-located with the master coordinator one.

Thank you in advance.

2

To answer the second part of your question:

 is there a way to configure a separate preview engine node?

You can always spin up an separate engine and assign the preview queries to go on it. Go to Settings → Queues → Edit “UI Previews” and change the engine from “preview” to any other engine.