2018-07-20 Integrations: Important enhancements to GCP and AWS monitoring

Released on: 
July 20, 2018

New for Amazon Web Services monitoring

  • New Relic has optimized the way to fetch metrics from AWS Cloudwatch. This has resulted in a significant reduction in the number of calls to your AWS Cloudwatch service, as well as a reduction in data lag. The list of specific permissions that are required to fetch monitoring data from AWS services has been updated accordingly, with the additional cloudwatch:GetMetricData permission.
    Recommendation: Use a custom policy only when you are unable to use the ReadOnlyAccessmanaged policy from AWS. If you use a custom policy, be sure to check that the new permission is included.
  • AWS Redshift integration now provides four new metrics to monitor the performance of multi-node clusters: QueriesCompletedPerSecond, WLMQueriesCompletedPerSecond, QueryDuration, and WLMQueryDuration. Check AWS Redshift integration for details.
  • If you use AWS API Gateway, you can now fine-tune the data gathered for this service by specifying the stages you want to monitor with the new Filter by stage name prefixes. Check Configure polling frequency and data collection for cloud integrations for details.

New for Google Cloud Platform monitoring

  • Google Kubernetes Engine integration is now available. If you're using Kubernetes v1.10.2 or later, now you can get visibility into containers, nodes, and pods. This cloud integration is complementary to New Relic's Kubernetes on-host integration.
  • You can now link your GCP projects to New Relic using a service account. If you had already used a user account to link your projects, you can migrate to the new authorization type, though the UI. Check Connect Google Cloud Platform services to Infrastructure for details.

Changes in Google Cloud Platform integrations

  • The default dashboards for Google Cloud Platform integrations now include a filter by Project ID. This is very useful if you monitor more than one project from the same Google Stackdriver Monitor service.