Separate out jobs count query to it's own network call

What does this MR do and why?

I abstracted the count from the jobs query to prevent timeouts. But it turned out that these network calls were still batched together. So when count query fails the query to get the jobs still failed as well due to a timeout.

This MR explicitly makes the count query it's own network call.

Screenshots or screen recordings

How to set up and validate locally

To reproduce edit app/graphql/types/limited_countable_connection_type.rb's count method to sleep(60) at the first line of the method.

MR acceptance checklist

This checklist encourages us to confirm any changes have been analyzed to reduce risks in quality, performance, reliability, security, and maintainability.

• I have evaluated the MR acceptance checklist for this MR.

Related to #389347 (closed)