Use cluster ID from the metrics to map to cluster name in the reports

[naved001]

All the metrics include a cluster ID. We should map the cluster ID to a cluster name and include that in a report.

@QuanMPhm For all other reports, where do we want to read the cluster name from? the file name or a value from a column?

[QuanMPhm]

@naved001 I can take over the issue. I will add the cluster names, either ocp-prod or ocp-test, into a column named Cluster Name. I'll submit a draft PR while waiting for your cluster name mapping.

[naved001]

@QuanMPhm so we'll need to first extract the cluster ID

I think in merge_metrics we could do something like

    def merge_metrics(self, metric_name, metric_list):
        """Merge metrics (cpu, memory, gpu) by pod"""
        for metric in metric_list:
            pod = metric["metric"]["pod"]
            namespace = metric["metric"]["namespace"]
            node = metric["metric"].get("node")
            cluster_id = metric["metric"].get("managed_cluster", "UnknownID")

            self.merged_data.setdefault(namespace, {})
            self.merged_data[namespace].setdefault(pod, {"metrics": {},
                                                         "cluster_id": cluster_id})
            # and the rest of the method continues

And when we are producing the reports, that's where we can read from a file that holds that mapping for cluster id to the names we use and put that in the report.

[QuanMPhm]

@naved001 I was thinking of just using a static dict in the code for now. For Openstack, I'm just adding a hardcoded string.

[naved001]

Change of plans. When collecting the metrics we'll write the URL as the cluster ID, and then we can map that to a friendly name when generating reports. This avoids us having to change the data structure we build for metrics.

{
"https://thanos-querier-openshift-monitoring.apps.shift.nerc.mghpcc.org": "nerc-ocp-prod",
"https://thanos-querier-openshift-monitoring.apps.ocp-test.nerc.mghpcc.org": "nerc-ocp-test",
}