New and updated features.

These release notes give an overview of the changes for Release 2020.10 of the VMware Aria Operations for Applications (previously known as Tanzu Observability by Wavefront) service. We also recently released a new version (6.1) of the Wavefront proxy. See the Wavefront Proxy Release Notes for details.

New Histogram Functions

This release includes the following new functions for managing histograms:

  • frequencyHisto() converts a histogram that comes from Google Cloud Platform, Telegraf, or any other source to Wavefront histogram format.
  • stddev() shows how the data in a histogram expression varies against the mean. Returns those values as a time series.

AWS Integration Changes

Trusted Adviser Service Limit Metrics

AWS Metrics+ Service Limit Metrics are now available in Operations for Applications.

For examples, see AWS Metrics+ Trusted Advisor Service Limits.

For sample alerts, see Alert Queries for AWS Trusted Advisor Service Limits

Changes to Required Permissions and CloudWatch API

If you are currently using an IAM Policy with Limited Access, review our changes to the Giving Limited Access documentation.

Changes support more efficient and new APIs we are using. For example, in the next release (2010.14.x) we will start querying Cloudwatch using cloudwatch:GetMetricData API calls (instead of the older cloudwatch:GetMetricStatistics API calls). As a result, we can fetch multiple metrics in one API call and significantly reduce the number of API calls for retrieving metrics from AWS.

Operators gt, lt, ge, le, eq, ne

The new operators allow you to easily filter on the query line. You can use multiple operators, for example, to find values that are between specified values, as in this example:

ts(~sample.mem.used.percentage).gt(70).lt(85)

~collector Metrics Changes

This release includes a more fine-grained set of ~collector metrics, for example, we now support:

  • ~collector.points.reported
  • ~collector.histograms.reported
  • ~collector.tracing.spans.reported
  • ~collector.tracing.span_logs.reported
  • ~collector.tracing.span_logs.bytes_reported

At the same time, we’ve removed the ~collector.*.valid metrics because the ~collector.*.reported metrics give more accurate results.