Controller metrics

[somnusfish]

Issue number:
#128

Description of changes:
Emitting prometheus metrics for host's bottlerocket version and host's brupop state.
Adding HTTP server in the controller to vend prometheus metrics.

Testing done:
Provisioned 5 old bottlerocket hosts in the cluster and run Brupop, curled the controller endpoint to verify the metrics.

Sample metrics during brupop updating hosts.

[ec2-user@ip-192-168-53-104 ~]$ curl 10.100.135.35/metrics
# HELP brupop_hosts_state Brupop host's state
# TYPE brupop_hosts_state gauge
brupop_hosts_state{state="Idle"} 6
brupop_hosts_state{state="MonitoringUpdate"} 1
brupop_hosts_state{state="PerformedUpdate"} 1
brupop_hosts_state{state="RebootedIntoUpdate"} 1
brupop_hosts_state{state="StagedUpdate"} 1
# HELP brupop_hosts_version Brupop host's bottlerocket version
# TYPE brupop_hosts_version gauge
brupop_hosts_version{bottlerocket_version="1.2.0"} 4
brupop_hosts_version{bottlerocket_version="1.5.1"} 2

Deployed prometheus on the cluster and use port forward to get he graphic analysis locally.

Terms of contribution:

By submitting this pull request, I agree that this contribution is dual-licensed under the terms of both the Apache License, version 2.0, and the MIT license.

[somnusfish]

As per discussion, update the metrics from controller spec change itself to monitoring the whole state for brupop cluster.