[Infra UI] Display cloud metrics on the node detail page

[skh]

Depends on #39280 and #39281

For AWS instances we would like to show metric data from the aws module, if available, together with other metricbeat data on the node detail page.

This task is to use the changed metadata and metrics endpoints on the node detail page to request and display cloud metrics when they are available.

This depends on #39280 and #39281 being implemented.

[elasticmachine]

Pinging @elastic/infra-logs-ui

[jasonrhodes]

@hbharding we should talk about this once we understand what metrics we are bringing back from the metrics query

[skh]

AWS metrics are documented here: https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/viewing_metrics_with_cloudwatch.html

[skh]

@hbharding @sorantis

FYI: these are the metrics that I see in AWS' CloudWatch interface:

and these are the metrics that we get through the aws.ec2 metricset:

"aws": {
      "ec2": {
        "diskio": {
          "write": {
            "count": 0,
            "bytes": 0
          },
          "read": {
            "count": 0,
            "bytes": 0
          }
        },
        "network": {
          "in": {
            "packets": 34.4,
            "bytes": 6359.8
          },
          "out": {
            "bytes": 133616,
            "packets": 69.2
          }
        },
        "status": {
          "check_failed_instance": 0,
          "check_failed_system": 0,
          "check_failed": 0
        },
        "instance": {
          "threads_per_core": 1,
          "public": {
            "ip": "18.218.163.78",
            "dns_name": "ec2-18-218-163-78.us-east-2.compute.amazonaws.com"
          },
          "private": {
            "dns_name": "ip-172-31-47-9.us-east-2.compute.internal",
            "ip": "172.31.47.9"
          },
          "image": {
            "id": "ami-0d8f6eb4f641ef691"
          },
          "state": {
            "name": "running",
            "code": 16
          },
          "monitoring": {
            "state": "disabled"
          },
          "core": {
            "count": 1
          }
        },
        "cpu": {
          "credit_usage": 0.023817,
          "credit_balance": 144,
          "surplus_credit_balance": 0,
          "surplus_credits_charged": 0,
          "total": {
            "pct": 0.43118458831162565
          }
        }
      }

[sorantis]

Looks aligned with the docs on EC2 metricset. Are we also looking into aws.tags.*? This could be useful for visualisation (e.g. filtering).

[skh]

Are we also looking into aws.tags.*? This could be useful for visualisation (e.g. filtering).

The node detail page is node-specific, i.e. there is not much to filter as we're already showing one specific instance, but we could definitely show the tags. We currently have no designs how to show non-timeseries / non-metric data on the node detail page in general, but that doesn't mean it's not possible.

[sorantis]

Agreed. Simply showing tags for a given instance can be a good start. If tags are clickable, they can be helpful for users to see what other resources have the same tag (redirecting to a respective Map/Table view). Generally tags have many applications.