Log results of cluster health check

[danielmitterdorfer]

When the cluster health check fails e.g. due to invalid setup where we
check for a green cluster on a single node but indices have been
configured with replicas, the benchmark appears to hang indefinitely.
The reason is that the cluster health check is retried by default but
there is no indication in the logs.

With this commit we log the result of the cluster health check. We also
raise the log level of retry log messages so users can analyze the
behavior by inspecting logs.

Closes #1150

[danielmitterdorfer]

Here are some examples of the behavior from my local testing:

In the success case we see in the logs:

PID:10002 esrally.driver.runner INFO cluster-health: expected status=[green], actual status=[green], relocating shards=[0], success=[True].

When the cluster health status is not reached, we get:

PID:11154 esrally.driver.runner INFO [cluster-health] has timed out. Retrying in [0.50] seconds.

and additionally, the Python client issues a warning right before the log line above:

PID:11154 elasticsearch WARNING GET http://127.0.0.1:39200/_cluster/health/logs-*?wait_for_status=green&wait_for_no_relocating_shards=true [status:408 request:30.006s]

[DJRickyB]

This is great, thank you. Upping the log level for retry is a good idea

[danielmitterdorfer]

Thanks for the review!