Add timeout attribute for NewRelic metrics provider

[orlando-valdez-ck]

Summary

Add an optional timeout attribute to the NewRelic provider spec in the AnalysisTemplate so we can send that to the NerdGraph Graphql API. This can look something like this,

apiVersion: argoproj.io/v1alpha1
kind: AnalysisTemplate
metadata:
  name: cpu-usage
  namespace: default
spec:
  metrics:
    - name: cpu-usage
      interval: 2m
      failureCondition: "result['percentile[95]'] >= 70"
      failureLimit: 4
      provider:
        newRelic:
          timeout: 10 # this timeout will be sent to and applied by the NerdGraph API to the query
          profile: newrelic
          query: "FROM K8sContainerSample SELECT percentile(`cpuCoresUtilization`, 95)['95'] SINCE 1 WEEK AGO"

When not configured, the default of 5 seconds can be sent to the API.

Use Cases

This can be used when using queries that take longer than the default of 5 seconds that the NerdGraph API uses when no timeout value is sent. This will avoid an AnalysisRun to error out with NRDB query duration exceeded the set timeout errors returned by the GraphQL API.

---

Message from the maintainers:

Impacted by this bug? Give it a 👍. We prioritize the issues with the most 👍.