Looker instance Slowness

Hi Everyone! I am hoping for some guidance on my Looker instances slowness as well as some best practices from fellow Ops folks who support Looker. 

Our looker instance is, and has been for some time, crazy slow. By slow I mean the UI is slow to render and even simple API calls take several seconds on occasion. I have verified that the architecture (VM hosted in Azure and RDS instance in AWS) appear to be underutilized and have very little latency.

We have what I believe to be an unusually high amount of objects in Looker. I was hoping you could let me know if this is normal:

  • 3,196 Dashboards
  • 7,923 Looks
  • 2,227 Scheduled Plans
  • 241 Explores
  • 14,464 Users (of which 11,009 are external standard users)


​​​​​Are these numbers far higher than recommended? Or is this normal?

Thanks in Advance

1 REPLY 1

Hi @jfritzz,

I don't know the recommended value, but I believe it has a particularly large number of users. I think a large number of Explore will affect LookML validation time more than API calls.

Check the Performance Recommendations dashboard for hints on how to improve performance.

Also, I have heard that you can scale up your instance by contacting Looker support. I think it would be a good idea to talk to them about it.
https://help.looker.com/hc/en-us/requests/new