High CPU usage on query services


We are testing out our server setup.

We have the following servers.

3 data servers
6 query services
2 index servers

We’ve noticed that the cbq-engine process on the query service servers are using 300% cpu usage without a reason. Looking at the admin ui, query service servers are processing only 3~6 n1ql queries per second. Most of our queries get timed out at this point.

How can I approach to debug/fix it?

We are on 4.1 enterprise with 3M+ documents and all of the queries are indexed.

Hi, moono326,

Would it be possible for you to share the index/query definition with us? This will help us better understand the issue. Data model will help too. Feel free to mask the data if they are sensitive. We just need to the following:

  1. Structure of the data
  2. data size.
  3. Number of documents per query
  4. SDK client etc.

You can also look into index/query logs:



I actually posted a new thread.

Is this related? Range() response failed `Index scan timed out`

Everything has been working fine for the last 4~7 days.

I will post 1~4 if that thread does not help.