Change log level to warning when temporarily partial EC2 info

Change log level to warning when temporarily unable to retrieve EC2 instance info from DescribeInstances call.

The from_describe_instance_data log a warning and throw an exception.
The from_describe_instance_data is called by
1. fleet_manager.launch_ec2_instances, but here the instance info are already populated in the _launch_instances, which calls the _get_instances_info that does have the retries, by calling _retrieve_instances_info_from_ec2
2. fleet_manager._retrieve_instances_info_from_ec2 that catch the exception

If at the end of the retires there are still instances info with partial data, in the _launch_instances these instances are finally logged as error

Signed-off-by: Luca Carrogu <[email protected]>

src/slurm_plugin/fleet_manager.py

@@ -61,7 +61,7 @@ def from_describe_instance_data(instance_info):
  instance_info["LaunchTime"],
  )
  except KeyError as e:
- logger.error("Unable to retrieve EC2 instance info: %s", e)
+ logger.warning("Unable to retrieve EC2 instance info: %s", e)
  raise e