implement metrics-aggregator parent chart, defined values

[jterzis]

Initial implementation of metrics-aggregator service using parent / sub chart Helm pattern.

The chart defined in Chart.yaml contains 3 subchart references:

1. datadog agent
2. prometheus
3. river-metrics-discovery

Configuration values in metrics-aggregator/values.yaml are only provided for datadog with the intent of matching
configuration with currently deployed agent in AWS metrics-aggregator ECS service.

To test the metrics-aggregator chart:

helm dependency update helm-charts/charts/metrics-aggregator
# view rendered manifest that will be deployed to cluster
cd charts && helm install metrics-aggregator ./metrics-aggregator --dry-run

In a subsequent PR, I will add prometheus and river-metrics-discovery with identical configuration as is currently deployed based on terraform config. I will also use make use of subchart template references using include to reuse templates from the parent chart in the subchart for river-metrics-discovery which will be defined with custom manifest files in this repo.

Update 01/15:

Had to make values changes for datadog, prometheus charts to deprecate unneeded images from the Deployment and allow for GKE Autopilot to work with datadog chart. As of now when running a helm install metrics-aggregator . on the cluster, datadog agent (as a daemonSet) and prometheus server are both created successfully running in the cluster.

Will add river-metrics-discovery service last as a custom subchart Deployment of the image in a ReplicaSet.

Note: We need to make sure we're okay with using the prometheus-community helm chart and associated prometheus image. It seems to be the best maintained one I could find.