Alert rule evaluation on Grafana Labs

No Data and Error states

Fri, 03 Apr 2026 19:43:06 +0000

No Data and Error states

Grafana Alerting implements the No Data and Error states to handle common scenarios when evaluating alert rules, and you can modify their behavior.

An alert instance can transition to these special states:

No Data state occurs when the alert rule query runs successfully but returns no data points.
Error state occurs when the alert rule fails to evaluate its query or queries successfully.

Note
No Data and Error states are supported only for Grafana-managed alert rules.

Tip
For common examples and practical guidance on handling Error, No Data, and stale alert scenarios, refer to the Handle connectivity errors and Handle missing data guides.

Alert instance states

A Grafana-managed alert instance can be in any of the following states, depending on the outcome of the alert rule evaluation:

State	Description
Normal	The state of an alert when the condition (threshold) is not met.
Pending	The state of an alert that has breached the threshold but for less than the pending period.
Alerting	The state of an alert that has breached the threshold for longer than the pending period.
Recovering	The state of a firing alert when the threshold is no longer breached, but for less than the keep firing for period.
Error^*	The state of an alert when an error or timeout occurred evaluating the alert rule. You can customize the behavior of the Error state, which by default triggers a different alert.
No Data^*	The state of an alert whose query returns no data or all values are null. You can customize the behavior of the No Data state, which by default triggers a different alert.

`Error` state

The Error state is triggered when the alert rule fails to evaluate its query or queries successfully.

This can occur due to evaluation timeouts (default: 30s) or repeated failures (default: 3) when querying the data source. The evaluation_timeout and max_attempts options control these settings.

When an alert instance enters the Error state, Grafana, by default, triggers a new DatasourceError alert. You can control this behavior based on the desired outcome of your alert rule in Modify the No Data or Error state.

`No Data` state

The No Data state occurs when the alert rule query runs successfully but returns no data points at all.

When an alert instance enters the No Data state, Grafana, by default, triggers a new DatasourceNoData alert. You can control this behavior based on the desired outcome of your alert rule in Modify the No Data or Error state.

Modify the `No Data` or `Error` state

These states are supported only for Grafana-managed alert rules.

In Configure no data and error handling, you can change the default behavior when the evaluation returns no data or an error. You can set the alert instance state to Alerting, Normal, Error, or Keep Last State.

Configure	Set alert state	Description
No Data	No Data	The default option for No Data events. Sets alert instance state to `No Data`. The alert rule immediately creates a new `DatasourceNoData` alert instance after evaluation, with the alert rule’s name, UID, and the data source UID as labels.
Error	Error	The default option for Error events. Sets alert instance state to `Error`. The alert rule immediately creates a new `DatasourceError` alert instance after evaluation, with the alert rule’s name, UID, and the data source UID as labels.
No Data or Error	Alerting	Sets the alert instance state to `Pending` and then transitions to `Alerting` once the pending period ends. If you sent the pending period to 0, the alert instance state is immediately set to `Alerting`.
No Data or Error	Normal	Sets alert instance state to `Normal`.
No Data or Error	Keep Last State	Maintains the alert instance in its last state. Useful for mitigating temporary issues.

Note that when you configure the No Data or Error behavior to Alerting or Normal, Grafana attempts to keep a stable set of fields under notification Values. If your query returns no data or an error, Grafana re-uses the latest known set of fields in Values, but will use -1 in place of the measured value.

Keep last state

The “Keep Last State” option helps mitigate temporary data source issues, preventing alerts from unintentionally firing, resolving, and re-firing.

However, in situations where strict monitoring is critical, relying solely on the “Keep Last State” option may not be appropriate. Instead, consider using an alternative or implementing additional alert rules to ensure that issues with prolonged data source disruptions are detected.

`No Data` and `Error` alerts

When an alert rule evaluation results in a No Data or Error state, Grafana Alerting immediately creates a new alert instance —skipping the pending period—with the following additional labels:

alertname: Either DatasourceNoData or DatasourceError depending on the state.
datasource_uid: The UID of the data source that caused the state.
rulename: The name of the alert rule that originated the alert.

Note that DatasourceNoData and DatasourceError alert instances are independent from the original alert instance. They have different labels, which means existing silences, mute timings, and notification policies applied to the original alert may not apply to them.

You can manage these alerts like regular ones by using their labels to apply actions such as adding a silence, routing via notification policies, and more.

If the alert rule is configured to send notifications directly to a selected contact point (instead of using notification policies), the DatasourceNoData and DatasourceError alerts are also sent to that contact point. Any additional notification settings defined in the alert rule, such as muting or grouping, are preserved.

Reduce `No Data` or `Error` alerts

To minimize the number of No Data or Error state alerts received, try the following.

Use the Keep last state option. For more information, refer to the section below. This option allows the alert to retain its last known state when there is no data available, rather than switching to a No Data state.
For No Data alerts, you can optimize your alert rule by expanding the time range of the query. However, if the time range is too big, it affects the performance of the query and can lead to errors due to timeout.

To minimize timeouts resulting in the Error state, reduce the time range to request less data every evaluation cycle.
To reduce multiple notifications from Error alerts, define a notification policy to handle all related alerts with alertname=DatasourceError, and filter and group errors from the same data source using the datasource_uid label.
Change the evaluation timeout (default: 30s) or the retry mechanism (max_attempts) settings. This should be a last resort, as it can affect the performance of all alert rules and may cause missed evaluations if the timeout is too long. For Grafana Cloud, open a support ticket from the Cloud Portal.

Tip
For common examples and practical guidance on handling Error, No Data, and stale alert scenarios, refer to the Handle connectivity errors and Handle missing data guides.

`grafana_state_reason` for troubleshooting

Occasionally, an alert instance may be in a state that isn’t immediately clear to everyone. For example:

If “no data” handling is configured to transition to a state other than No Data.
If “error” handling is configured to transition to a state other than Error.
If the alert rule is deleted, paused, or updated in some cases, the alert instance also transitions to the Normal state.
Stale alert instances in the Alerting state transition to the Normal state when the series disappear.

In these situations, the evaluation state may differ from the alert state, and it might be necessary to understand the reason for being in that state when receiving the notification.

The grafana_state_reason annotation is included in these situations, providing the reason that explains why the alert instance transitioned to its current state. For example:

If “no data” or “error” handling transitions to the Normal state, the grafana_state_reason annotation is included with the value No Data or Error, respectively.
If the alert rule is deleted or paused, the grafana_state_reason is set to Paused or RuleDeleted. For some updates, it is set to Updated.
Stale alert instances in the Normal state include the grafana_state_reason annotation with the value MissingSeries.

Stale alert instances

Fri, 03 Apr 2026 19:43:06 +0000

Stale alert instances

An alert instance is considered stale if the alert rule query returns data but its dimension (or series) has disappeared for a number of consecutive evaluation intervals (2 by default).

This is different from the No Data state, which occurs when the alert rule query runs successfully but returns no dimensions (or series) at all.

A stale alert instance transitions to the Normal (MissingSeries) state as Resolved, and is then evicted:

Eval. Interval	1	2	3	4
Alert instance A	✔	✔	✔	✔
Alert instance B	✔	`MissingSeries`	️`Normal(MissingSeries)` 📩^*

Note
Stale alert instances are supported only for Grafana-managed alert rules.

How Grafana handles stale alert instances

The process for handling stale alert instances is as follows:

The alert rule runs and returns data for some label sets.
An alert instance that previously existed is now missing.
Grafana keeps the previous state of the alert instance for the number of evaluation intervals specified in Missing series evaluations to resolve.
If it remains missing after the specified number of evaluation intervals (2 by default), it transitions to the Normal state and sets MissingSeries in the grafana_state_reason annotation.

Stale alert instances in the Alerting, No Data, or Error states transition to the Normal state as Resolved, and are routed for notifications like other resolved alerts.
The alert instance is removed from the UI.

Tip
For common examples and practical guidance on handling No Data and stale alert scenarios, see Handling missing data.

Configure Missing series evaluations to resolve

In Configure no data and error handling > Missing series evaluations to resolve, you can set how many consecutive evaluation intervals must pass without data for a given dimension before the alert instance is marked as stale and resolved.

If you don’t specify a value, Grafana uses the default of 2 evaluation intervals.

Alert rule state and health

Fri, 03 Apr 2026 19:43:06 +0000

Alert rule state and health

Each alert rule can generate one or more alert instances—one alert instance for each series or dimension, as shown in the multi-dimensional alert example.

Each alert instance of the same alert rule represents a different target and can be in a different state; for example, one alert instance may be Normal while another is Alerting.

The alert rule state and alert rule health are determined by the state of the alert instances.

Alert rule states

An alert rule can be in either of the following states:

State	Description
Normal	None of the alert instances returned by the evaluation engine is in a `Pending` or `Alerting` state.
Pending	At least one alert instances returned by the evaluation engine is `Pending`.
Firing	At least one alert instances returned by the evaluation engine is `Alerting`.

Alert rule health

An alert rule can have one of the following health statuses:

State	Description
Ok	No error when evaluating the alert rule.
Error	An error occurred when evaluating the alert rule.
No Data	The alert rule query returns no data.
{status}, KeepLast	The rule would have received another status but was configured to keep the last state of the alert rule.

How rules are evaluated within a group

Fri, 03 Apr 2026 19:43:06 +0000

How rules are evaluated within a group

Each evaluation group contains an evaluation interval that determines how frequently the rule is evaluated. For instance, the evaluation may occur every 10s, 30s, 1m, 10m, etc.

Rules in different evaluation groups can be evaluated simultaneously.

Rules within the same evaluation group can be evaluated simultaneously or sequentially, depending on the rule type:

Grafana-managed rules within the same group are evaluated concurrently—they are evaluated at different times over the same evaluation interval but display the same evaluation timestamp.
Data source-managed rules within the same group are evaluated sequentially, one after the other—this is useful to ensure that recording rules are evaluated before alert rules.
Grafana-managed rules imported from data source-managed rules are also evaluated sequentially.

Alert rule evaluation on Grafana Labs

No Data and Error states

No Data and Error states

Alert instance states

Error state

No Data state

Modify the No Data or Error state

Keep last state

No Data and Error alerts

Reduce No Data or Error alerts

grafana_state_reason for troubleshooting

Stale alert instances

Stale alert instances

How Grafana handles stale alert instances

Configure Missing series evaluations to resolve

Alert rule state and health

Alert rule state and health

Alert rule states

Alert rule health

How rules are evaluated within a group

How rules are evaluated within a group

`Error` state

`No Data` state

Modify the `No Data` or `Error` state

`No Data` and `Error` alerts

Reduce `No Data` or `Error` alerts

`grafana_state_reason` for troubleshooting