Soperator Exporter: add node metrics.

[theyoprst]

➜ curl localhost:8080/metrics
# HELP slurm_active_node_gpu_seconds_total Total GPU seconds on active Slurm nodes (not down, not idle+drain)
# TYPE slurm_active_node_gpu_seconds_total counter
slurm_active_node_gpu_seconds_total{node_name="worker-0"} 30.072787542
slurm_active_node_gpu_seconds_total{node_name="worker-1"} 51.23264987500001
# HELP slurm_node_fails_total Total number of times a node has failed (went from not down/drain to down/drain state)
# TYPE slurm_node_fails_total counter
slurm_node_fails_total{node_name="worker-0",reason="Scheduled maintenance"} 2
# HELP slurm_node_info Slurm node info
# TYPE slurm_node_info gauge
slurm_node_info{address="10.0.15.254",base_state="idle",compute_instance_id="computeinstance-xxx",is_drain="true",node_name="worker-0"} 1
slurm_node_info{address="10.0.63.138",base_state="idle",compute_instance_id="computeinstance-yyy",is_drain="false",node_name="worker-1"} 1
# HELP soperator_cluster_info Soperator cluster information
# TYPE soperator_cluster_info gauge
soperator_cluster_info{soperator_version="1.19.xxx"} 1