Understand metrics structure and how to explore metrics in the Metrics Browser

VMware Aria Operations for Applications (formerly known as Tanzu Observability by Wavefront) provides observability for several different metric types including time series metrics, delta counters, histograms, and traces/spans. This page looks at the anatomy of a time series metric and shows you how to explore it in the Metrics Browser.

## Videos

The following videos get you started. Note that these videos were created in 2020 and some of the information in them might have changed. They also use the 2020 version of the UI.

 Browsing Your Data 90-second video that shows how you can find and examine metrics from the Sources browser and from the Metrics browser. You can also watch the video here . About Cardinality Wavefront co-founder Clement Pang explains why the concept of cardinality is so important for observability, what high cardinality means, and why we deal so well with high cardinality input. You can also watch the video here .

## Time Series Metric Structure

A time series has, at a minimum, the metric name, value/timestamp, and source. In many cases, the metric is ingested with additional information represented as tags.

### Simple Time Series

Here’s one example that shows the minimum elements of a time series.

Each time series is a unique combination of:

• Metric name–Describes the metric. There’s often a hierarchy of metrics, each with a corresponding time series.
• Value & Timestamp–Value at the specified time.
• Source–The source of the metric. Host, VM, etc. In contrast to some other observability platforms, this dimension is always part of the metric.

Here’s a screenshot of the time series that is shown in the diagram above in a chart.

### Time Series with Tags

In most cases, the time series includes one or more tags to allow a more fine-grained analysis. The `~sample` data you can find on each service instance include point tags for environment and availability zone.

Point tags offer a powerful way of labeling data so that you can slice and dice it in almost any way you can imagine. For example, you can use point tags, to label a point’s datacenter, version, etc. and can then group by datacenter or version.

You use point tags to add extra dimensions to your data, and can then focus your exploration just on that dimension.Fine Tune Queries with Point Tags explains how to use point tags.

Here’s a screenshot of the time series that includes point tags in a chart.

### How Filtering with Tags Improves Usability

How the point tag filters are useful becomes obvious when the `source=` filter is removed. The result of all time series for `~sample.disk.space.used` is visually confusing.

When you add filters for `env` and `az`, the information makes sense.

## Obsolete Metrics

If a metric stops sending data points for a certain period of time (obsolescence period), it becomes obsolete.

In the Metrics browser and Query Editor, obsolete metrics are no longer shown in the autocomplete drop-down lists.

## Metrics Browser

Select Browse > Metrics to display the Metrics Browser. Use the Metrics Browser to find metrics that are actively sending data points.

On the Metrics Browser, you can:

• Drill down and go up the hierarchy.
• Filter by name or source.
• Hide and redisplay individual metrics or metrics namespaces to unclutter your page.
• View the metric type in terms of retention period - persistent or ephemeral.
• Convert persistent metrics to ephemeral and the reverse. Most of the internal metrics are ephemeral and not convertible.
• Create a chart or dashboard for an individual metric or for the current set of metrics.
• View the sources and point tags for an individual metric.

### Hide and Redisplay Metrics

While obsolete metrics are automatically hidden, you can manually hide metrics from the Metrics browser. Manually hiding metrics does not permanently delete a metric or metric namespace.

 To hide one or more metrics: Select Browse > Metrics. Click the Manage Hidden Metrics button. In the dialog box, type a complete metric name (e.g. `requests.latency`) or a metric prefix (e.g. `requests.`, `cpu.loadavg.`). This field does not support auto-complete, so you have to type the entire metric name or metric prefix. The text is case sensitive. Wildcards are not supported. The star `*` character is considered part of the text string. Press Enter to add the metric(s) to the list and click Save.
 To view hidden metrics: Select Browse > Metrics. Click the Manage Hidden Metrics button. Click the Unhide button to the right of the metric or metric prefix to unhide and click Save. The selected metrics and metric prefixes appear again as long as they are not obsolete.

### Change the Retention Period of Metrics

With the 2024-05 release, we introduce ephemeral metrics, which have short retention period. By default, all ingested metrics are persistent but are convertible to ephemeral.

Converting persistent metrics to ephemeral can significantly improve the query performance and reduce the cardinality.

 Select Browse > Metrics. Click the Change Ephemerality button. To convert a persistent metric or metrics namespace to ephemeral, in the Select Metric Prefix text box, enter the target metric name or namespace prefix, and press Enter. The metric name or namespace prefix appears in the Ephemeral Metrics table below. You can repeat this step for multiple metrics and metrics namespaces. To convert an ephemeral metric or metrics namespace to persistent, in the Ephemeral Metrics table, locate the target metric or namespace prefix and click the corresponding Convert to Persistent Metric action.The metric name or namespace prefix disappears from the Ephemeral Metrics table. You can repeat this step for multiple metrics and metrics namespaces. Click Save.

Changing the retention period of a metric or metrics namespace creates a System event:

• Converting a persistent metric to ephemeral creates a System event with the name `Ephemeral Prefix: Added <metric_name>`.
• Converting an ephemeral metric to persistent creates a System event with the name `Ephemeral Prefix: Removed <metric_name>`.