Move all loops based on retries to time wait

Nobody has claimed this yet.

It is hard to calculate how many time we will wait for each loop when they are based on counting retries

Loops like:

```
for trying in range(1, 101):
[...]
logging.error("Failed to get namespace for %s on the %s cluster after 100 retries" % (cluster_name, type))
```

Can be moved to:

```
wait_time = 30 # (minutes)
starting_time = datetime.datetime.utcnow().timestamp()
while datetime.datetime.utcnow().timestamp() < starting_time + wait_time * 60:
[...]
logging.error("Waiting time expired. After %d minutes there are %d ready nodes (Expected: %d) on cluster %s" % (wait_time, ready_nodes, worker_nodes, cluster_name))
```

No files, tests, or entry points are named. Locate the Python loops that use retry counts, read their surrounding cluster-wait logic, and identify the intended timeout for each; done means those loops use elapsed waiting time and report the timeout and current state consistently.

Written by the indexing model from the issue text.