Documentationbreadcrumb arrow Pluginsbreadcrumb arrow Jira data sourcebreadcrumb arrow Alerting
Grafana Cloud Enterprise
Last reviewed: September 2, 2026

Set up alerting for Jira data

The Jira data source supports Grafana Alerting. Alert evaluation runs on the backend. Use alerts to get notified when issue counts exceed a threshold, work is overdue, or an SLA has breached.

Before you begin

Query requirements

Jira queries return a table: one row per issue, not a Prometheus-style time series. Grafana Alerting’s default Reduce and Threshold expressions expect a wide time-series frame and fail on that table with an error such as input data must be a wide series.

To evaluate Jira results:

  1. Select at least one field in Select Fields. For counts, Key is enough. For sums, select a numeric field such as Story point estimate.
  2. Filter with JQL. Don’t use dashboard template variables such as $project. Alert rules evaluate outside the dashboard, so those variables are empty or wrong. Hard-code project keys, or use $__timeFrom and $__timeTo (they follow the alert rule time range).
  3. Set Limit higher than the number of issues you expect. The default is 50. If Limit is lower than the matching issue count, the alert under-counts.
  4. Add a Classic condition on query A (for example, WHEN count() OF A IS ABOVE 50). Use count() for issue counts and sum() for numeric fields.

Queries that return only text, such as Summary with no aggregation condition, can’t be evaluated as a threshold.

Create an alert rule

To create an alert rule based on Jira data:

  1. Go to Alerts & IRM > Alert rules.
  2. Click New alert rule.
  3. Enter a name for the rule.
  4. In Define query and alert condition:
    • Select your Jira data source.
    • Select fields and enter JQL. Refer to Jira query editor.
    • Add a Classic condition that uses count() or sum() on query A, then set the threshold.
  5. In Add folder and labels, choose a folder, and add labels and annotations so notifications include the project and a link to a dashboard.
  6. In Set evaluation behavior, choose an evaluation group, set the evaluation interval, and set the pending period (how long the condition must be true before the alert fires).
  7. Select a contact point, or configure contact points later.
  8. Click Preview to confirm the query runs, then click Save rule.

You can also open a dashboard panel, click Edit, open the Alert tab, and click Create alert rule from this panel. Copy the JQL into the rule and add a Classic condition. Panel transformations aren’t a substitute for that condition.

For detailed Grafana UI steps, refer to Create Grafana-managed alert rules.

Examples

The following examples assume a Classic condition on query A. Replace 'Your Project' with your project key and raise Limit so every matching issue is included.

Alert when too many issues are open

  1. Select Fields: Key
  2. Filter (JQL): project = 'Your Project' AND status != done
  3. Classic condition: WHEN count() OF A IS ABOVE 50
  4. Evaluation interval: 10m. Pending period: 10m.

Alert on overdue issues

  1. Select Fields: Key, Due Date
  2. Filter (JQL): project = 'Your Project' AND due < now() AND status != done
  3. Classic condition: WHEN count() OF A IS ABOVE 0

Alert on remaining story points in the current sprint

  1. Select Fields: Story point estimate
  2. Filter (JQL): project = 'Your Project' AND sprint in openSprints() AND type != epic AND status != done
  3. Classic condition: WHEN sum() OF A IS ABOVE 20 (use your team’s threshold)

Alert on new critical issues

  1. Select Fields: Key
  2. Filter (JQL): project = 'Your Project' AND priority IN ('Critical', 'Blocker') AND created >= -1h
  3. Classic condition: WHEN count() OF A IS ABOVE 0
  4. Evaluation interval: 5m. Pending period: 0s (or a short pending period if you want to avoid single-evaluation noise).

Alert on a breached SLA

Jira Query Language can filter SLA fields directly. You don’t need to parse the SLA JSON for this alert.

  1. Select Fields: Key
  2. Filter (JQL): project = 'Your Project' AND "Time to resolution" = breached() AND status != done
  3. Classic condition: WHEN count() OF A IS ABOVE 0

The SLA field name is instance-specific. Check Select Fields or Jira’s advanced search for the exact name (for example, Time to first response).

Best practices

  • Evaluation interval: Jira issue data usually changes slowly. Evaluate every 5 to 15 minutes unless you need faster detection for critical issues.
  • Pending period: Require the condition to stay true for one or more intervals so a single slow query doesn’t flap the alert.
  • Limit: Set Limit high enough that the count or sum includes every matching issue.
  • JQL in the rule: Hard-code project keys and statuses. Don’t rely on dashboard variables.
  • Preview: Use Preview on the rule before you enable notifications.
  • Contact points: Configure contact points (email, Slack, PagerDuty, and others) rather than legacy notification channels.

If Preview fails with input data must be a wide series, the rule is still using Reduce and Threshold on a Jira table. Switch the condition to a Classic condition with count() or sum().