Azure Monitor Metrics Usage Insights Reaches GA With Ingestion Volume Change Dashboard

What the feature is​

Azure Monitor workspaces are the store for Prometheus metrics, which are widely used to monitor containers and Kubernetes. Metrics volume tends to climb as cloud-native estates grow, and Redmondmag's report notes that this makes observability costs harder to control.

Microsoft Learn describes Metrics usage insights as covering four areas:

• time series and event usage
• throttling limits
• usage trends
• unused metrics

Microsoft's stated aim is to let you act on that data, for example by removing unused metrics and right-sizing resources.

The feature isn't new. Microsoft announced the public preview as aimed at Azure Managed Prometheus users, analyzing all metrics ingested into the workspace. The GA announcement adds the Ingestion Volume Change dashboard.

A note on labeling: the Microsoft Learn page I could read still carries a "Preview" label in its title. That is documentation lag, not evidence that the feature is still in preview. The GA announcement is the authoritative signal.

The new Ingestion Volume Change dashboard​

According to Microsoft Learn, this dashboard compares metric ingestion between two points in time. You use it to spot unexpected spikes or drops in time series counts and event ingestion rates. Microsoft says it can help you investigate cost changes and plan capacity. You can filter by namespace and metric.

How it works in practice:

1. Pick a range in the Comparison dates dropdown.
2. The dashboard compares the earliest and latest data available within that range. Microsoft warns this may differ from the exact dates you chose.
3. Use the Summary view for starting values and changes. Switch to Detailed to add ending values and daily event counts.

Microsoft's tip is to use wider ranges, such as 60 or 90 days, for long-term trends. Use shorter ranges to investigate recent changes.

The comparison table shows, for each metric:

• start and end time series counts
• the difference between them
• start and end event rates per minute
• start and end daily events

Two reading notes from the documentation:

• The percentage columns show magnitude only. Check the difference column to see whether the value rose or fell.
• If a metric's starting value was zero and its end value is not, the percentage shows 100%. That means the metric is entirely new. If both values are zero, the percentage is blank.

The other dashboards​

The Ingestion Volume Change dashboard sits alongside the existing views.

• Limits & usage shows time series and event ingestion against their limits. It also shows the percentage of each limit used. Microsoft ties these limits to throttling.
• Workspace exploration breaks usage down per metric, including namespace, dimensions, daily time series, event rate and daily events. Microsoft says this data is calculated once per day and is current as of the previous day. Treat it as an analysis view, not live telemetry.
• Top 10 metrics by time series growth compares each metric's 7-day average against its 28-day average. Microsoft suggests checking sampling and scraping frequencies if you get throttling alerts.
• Unused metrics lists metrics that haven't appeared in a query within the period you set in Days Since Last Query.

How to turn it on​

Insights depends on a diagnostic setting that sends usage data to a Log Analytics workspace.

New workspace: select Enable insights on the Monitoring tab of the create page and specify a Log Analytics workspace. The diagnostic setting is created for you.

Existing workspace (portal):

1. Open the Azure Monitor workspace and go to Monitoring → Diagnostic settings.
2. Select Add diagnostic setting.
3. Give it a descriptive name.
4. Select Send to Log Analytics workspace and choose the destination.
5. Select Metrics Usage Details, then save.

Microsoft also provides ARM and Bicep examples. They enable the MetricsUsageDetails log category on a Microsoft.Insights/diagnosticSettings resource scoped to the workspace.

Check it worked: wait about 24 hours. Then confirm that Active Time Series (Max) is not zero. A zero means nothing is being ingested into the Azure Monitor workspace, so there is nothing to show. If your dashboards look empty on day one, check this before anything else.

To see all workspaces in a subscription, open Insights Hub from the Monitor menu and select Azure Monitor workspaces. Each workspace's row shows its limits, active time series, events per minute and utilization percentages.

Cost, data and scope​

• Cost of the feature: Microsoft Learn says there is no additional charge for the usage data sent to Log Analytics, the queries or the storage. That covers the Insights data only. It does not mean Prometheus ingestion or other Azure Monitor and Log Analytics usage is free.
• Regions: Microsoft says the feature is available in all public regions where Azure Monitor workspace is GA.
• Raw data: usage data lands in the AMWMetricsUsageDetails table. You can query it directly for history beyond what the Insights pages show. You can also customize the underlying queries by editing the Metrics Usage Insights workbook.
• Consolidation: Microsoft suggests sending data from several Azure Monitor workspaces to one Log Analytics workspace for a combined view.

What it won't do​

The tool shows you the data and doesn't change your collection. Cutting cost still means editing Prometheus scrape-job settings or metrics relabeling. Microsoft's guidance points to those as the way to stop collecting unused data.

A metric showing as unused only means nobody queried it in the window you set. It may still back an alert rule, a quarterly report or an incident runbook. Confirm ownership and operational need before dropping anything. The Redmondmag summary does not give pricing rates or new service limits, so treat any savings estimate as your own calculation.

A sensible workflow​

1. Enable the diagnostic setting and wait a day.
2. Check Limits & usage to see how close you are to throttling.
3. Use Ingestion Volume Change with a short window to date a sudden shift, or 60–90 days for a trend.
4. Use Workspace exploration to find the metrics responsible.
5. Review scrape intervals and relabeling for those metrics. Then check Unused metrics for removal candidates, after confirming with the owning teams.

This workflow is a synthesis of Microsoft's documented steps, not an official runbook.

Bottom line​

This is not a flashy release, but it is useful for teams whose Prometheus bill or throttling alerts keep growing. The dashboard cannot make decisions for you. It does make it much quicker to see when ingestion jumped, and which metric caused it.