Fixed TestSchedulerEnqueueWithFrontendDisconnect flakyness

[pracucci]

What this PR does:
The test TestSchedulerEnqueueWithFrontendDisconnect is flaky (CI failed here). The reason is that there's no guarantee that the query-frontend disconnected from the scheduler and its request canceled before verifyQuerierDoesntReceiveRequest is called. The change done in this test should help to reduce this likelihood.

I also took the opportunity to:

1. Rename the metric cortex_query_scheduler_connected_workers to cortex_query_scheduler_connected_querier_clients, because:

• Now it's more clear which workers we're talking about
• In the frontend v1 the metric is called _clients, so I've renamed for consistency

2. Add the metric cortex_query_scheduler_connected_frontend_clients

Which issue(s) this PR fixes:
Fixes #

Checklist

• Tests updated
• Documentation added
• CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX]

[pstibrany]

LGTM, thanks!