Control metrics cost in the TCO Optimizer
Metrics in the TCO Optimizer is available to accounts with early access. Contact your Coralogix account representative to request it. Without it, you block and unblock metrics from Metric usage analysis.
The Metrics page of the TCO Optimizer shows how much each metric and APM service contributes to your metrics usage, and lets you block the costly ones from the same screen you use for logs, traces, and RUM. Each metric or APM service you block becomes its own block policy, so you can see at a glance what is blocked and remove a block when you need the data again.
What you need
- Early access to Metrics in the TCO Optimizer for your account.
METRICS.TCO:READPOLICIESto view the Metrics page, its policies, and its statistics.METRICS.TCO:UPDATEPOLICIESto create and delete block policies and to change dynamic rules.
For the roles that include these permissions, see Permissions.
Open the Metrics page
- Navigate to Data Flow, then TCO Optimizer.
- In the left navigation, select Metrics.
The page has two views:
- Metric policies: the dynamic rules and block policies in force. The header shows the number of block policies.
- Statistics: usage by metric or APM service for one day, where you can select rows and turn them into block policies.
The Filters section in the left navigation narrows both views by Metric, Service, and Priority.
Find costly metrics in Statistics
Select Statistics to explore usage by metric and APM service. The data is aggregated over one UTC day.
- Use the Metrics and APM services switch to choose what each row represents.
- Use the day picker to select the day. The arrows move to the previous or next day.
- Search by metric, APM service, or policy name.
- Under Group by, select Priority to group rows by the priority their data is stored at.
The table includes these columns. Hover the info icon in a column header for its definition.
| Column | What it shows |
|---|---|
| Metric or APM service | The metric name as ingested, or the APM service that emits the metrics. |
| Affecting policy | The block policy that applies to the row, Adaptive rules when a dynamic rule sets its priority, or None when nothing acts on it. Select a policy name to open it in Metric policies. |
| Priority | The priority the data is stored at, or High when nothing acts on it. |
| Usage and % Usage | The volume of metrics data sent on the selected day, and its share of total metrics usage. |
| Samples and % Samples | The metric data points ingested on the selected day, and their share of the total. |
| Dimensions and Variations | For metrics only: the count of unique label names, and the count of unique label combinations. |
| Cardinality and % Cardinality | For metrics: the count of unique time series, an approximation with a 1% error margin, and its share of the total. For APM services, these columns are Time series and % Time series. |
Rows are sorted by Usage, highest first, so the most costly metrics appear at the top.
Block a metric or APM service
You can block from Statistics or start from Metric policies.
Block from Statistics
- In Statistics, select the checkbox of each metric or APM service you want to block. You can select up to 50 rows at a time. Rows that are already blocked can't be selected, and their tooltip names the policy that blocks them.
- Select Create policy. The New metric policy panel opens with your selection in Filters.
- Review the Preview and select Create.
Block from Metric policies
- In Metric policies, select Create policy.
- In Filters, select
metric.nameorservice.name, then select or type the names to block. The list suggests names from your usage and skips names that are already blocked. You can add up to 50 names. - Review the Preview and select Create.
What the New metric policy panel contains
- Quota label: the panel header shows the units of your daily quota available to metrics, out of the whole daily quota, for example Metrics · Unallocated pool 2,809 / 4,280. Unallocated pool appears only when metrics have no quota rule. Hover the info icon for a card that names the pool, Reserved for metrics or Unallocated pool, with its units, a sentence explaining why that pool applies, one row per slice of the quota bar with the pool in use highlighted, and a Quota Rules link to reserve units for metrics.
- Details: the policy name is filled in for you. Each policy is named after the metric or service it blocks.
- Filters: one condition on
metric.nameorservice.name, matched with Is. Each selected metric or service is saved as its own policy. - Action: Block is the only action. Route & Store is shown but can't be selected, because metric policies can only block for now.
- Preview: the impact over the Last 7 days. The Overview tab shows Total data sent, % Usage, and the Data sent over time and Units over time charts. The Breakdown tab lists Metric or APM service, Usage, % Usage, and Samples for each selected name.
When you select several names, the button reads Create N policies, and a confirmation lists the policies that will be created, one per name.
The new policies appear in the Policies table.
What blocking does
A block policy drops the matched metrics at ingestion: they are not stored and not forwarded to any destination. Queries, dashboards, and alerts that read a blocked metric stop receiving new data for it.
Metric policies don't set quota-based priorities, and the Overview's Daily team quota and Daily quota left cards don't include metrics. Depending on your plan, data sent for a blocked metric can still count toward your usage at a reduced rate. To confirm how blocked metrics are counted for your account, contact your Coralogix account representative.
Manage block policies
The Policies table in Metric policies lists one policy per blocked metric or APM service, with its Policy name, Filters, Action, Destinations, and Priority. Search the table by policy name, metric.name, or service.name.
Block policies can't be edited, disabled, or reordered. To stop blocking a metric or APM service, delete its policy:
- In the Policies table, select the Delete policy icon on the policy row.
- Confirm with Delete. The metric or service is ingested again.
Dynamic rules
When adaptive rules are enabled for your account, Metric policies also shows a Dynamic rules table. These are predefined rules that act on every metric based on how it is used:
| Rule | What it matches |
|---|---|
| Unused (not queried in 14 days) | Metrics that no query has read in the last 14 days. |
| Bogus (queries never yield series) | Metrics whose queries never return a series. |
| Only broken queries (no success) | Metrics that are queried, but never successfully. |
| Ad-hoc only (no dashboard/alert/recording/SLO/API) | Metrics read only by ad-hoc queries. No dashboard, alert, recording rule, SLO, or API uses them. |
The rules apply from top to bottom, and the first matching rule sets the priority. You can't add, remove, or reorder them. With METRICS.TCO:UPDATEPOLICIES, you can set a rule's Priority to Block, Low, or Medium, and turn it on or off with the Enabled toggle. When a block policy and a dynamic rule both match a metric, the block policy takes precedence.
Metrics on the Overview page
With Metrics enabled, the Overview page adds a Metrics chart split by priority next to the Logs, Traces, and RUM charts, and Feature set by priority gets a Metrics tab. Usage by priority, Data blocked today, and Yearly savings include metrics volume.
Metric Data page
With Metrics enabled, you block and unblock metrics and APM services in the TCO Optimizer. On the Metric Data page, the block and unblock actions are no longer offered, blocked status is still shown, and a banner links to the TCO Optimizer.