Could you provide some more information about your cloud, for example:
- The number of requests that it handles per minute
- How much data you are indexing
- If there is any memory pressure
The ping handler merely sends a query to the collection and makes sure that
it responds healthily. Can you check your schema to see what the query is
that it is sending? The ping query may be more expensive than you think.
Also is the ping using distrib=true or false in the query?