add semaphore state to noobaa endpoint metrics

[naveenpaul1]

Explain the changes

1. This code change will add both namespace store and NSFS semaphore state to noobaa endpoint metrics s3-service-monitor

Issues: Fixed #xxx / Gap #xxx

1. semaphore metrics will give more understanding of semaphore related issues

Testing Instructions:

1. read/write data to either namespace store or NSFS S3 bucket.
2. verify Prometheus metrics, noobaa endpoint metrics should have the following metrics
   semaphore_value, semaphore_waiting_value, semaphore_waiting_time, semaphore_waiting_queue
   

• Doc added/updated
• Tests added

[dannyzaken]

a few general comments:

1. is there a reason to share the same metric name for object_io and namespace_fs semaphores? We can have both NSFS and noobaa bucket workloads in the same system. @nimrod-becker I remember something about reducing the number of metrics we send to Prometheus. Is this something to consider?
2. the metrics contain only one sample of the semaphore values taken only when there is IO. So once IO stops, the value will not be updated, am I right?
3. I think it can be more consistent with sampling the semaphores values in constant intervals (a function running in the background) and providing an average (e.g., for the last minute).

[liranmauda]

Have some comments