Skip to main content
The active alerts page provides a consolidated view of all currently firing alerts, helping you quickly understand the overall alert status of your system. You can filter and browse by severity, title, labels, and more.
The active alerts feature requires alert engine v0.64.0 or later. If you have not installed it yet, go to Alert engine management to deploy it.

Viewing active alerts

Menu entry: Select a folder, then switch to the “Active Alerts” tab The active alerts list displays currently firing alerts in a table format. Default columns include:

Filtering and searching

The conditions bar at the top of the page helps you quickly locate the alerts you care about:
  • Alert rule: Filter by the rule that produced the alert (always visible; supports searching by rule name)
  • Severity: Filter by Critical, Warning, or Info
  • Title: Filter by the title of the alert event
  • Hash: Filter by the unique identifier hash of the alert
  • Job: Filter by the job label value
  • Instance: Filter by the instance label value
  • Extra labels: Filter by other additional labels
All filter conditions use an “AND” relationship — only alerts matching all conditions are displayed. In the alert rules list, the Status column shows the rule’s runtime status and clicking it opens the runtime details drawer; the Active alerts column next to it is how many active alerts the rule currently has, and clicking that number switches you to the “Active Alerts” tab filtered by that rule (URL parameter active_alert_rule_id).

Customizing columns

Click the column configuration button in the upper-right corner of the list to:
  • Show/hide columns: Check or uncheck columns to control which ones are displayed
  • Reorder columns: Drag and drop to adjust the display order of columns
The system remembers your column display preferences and automatically applies them on your next visit.
When you add a label-type column, the corresponding label filter automatically appears in the conditions bar, making it easy to quickly filter by that label.

Viewing alert details

Each row has a View details entry, the eye icon at the end of the row. It opens the active alert detail drawer, which has the following sections:
  • Alert overview: current status (active), severity, active duration, and notification count
  • Investigation entries: the investigation entries snapshot of the rule that produced this alert, one row each, clickable through to the dashboard or the Query Workbench. This block shows the snapshot taken when the alert was produced, not the rule’s current configuration. See Investigation entries for snapshot semantics and what happens when rendering fails.
  • Time information: first fired, last evaluated, last notified, and last updated
  • Rule & source: alert rule, check mode (Threshold / Data exists / No data), data source name, data source type, instance type, and query Ref
  • Current values: a snapshot of the query result values that triggered the alert
  • Alert dimensions: the alert’s deduplication dimensions
  • Event labels: the labels carried by the alert event
  • Alert annotations: the annotations carried by the alert event
  • Notes description: the rule’s notes description, if present
  • Diagnostic information (collapsed): alert hash, lifecycle ID, rule ID, data source ID, alert engine cluster, alert engine instance ID, and rule version
Labels and values are displayed as key=value and can be copied with a click, which is handy when troubleshooting.

Closing alerts

If certain active alerts no longer need attention (for example, the alert rule has been adjusted but old alerts have not yet automatically recovered), you can close them manually:
1

Select alerts

Check the alerts you want to close in the list. Multiple selection is supported.
2

Submit the close request

Click the Close selected active alerts button. After confirming, the selected alerts will be removed from the active alerts list.
Closing only removes the alert from the active alerts list. If the alert condition is still met, the alert will reappear the next time the alert engine runs its check.

New alert objects per rule are rate limited

A rule can create only so many new alert objects per minute (configured platform-wide, 1000 by default). Past that limit, the rule’s extra firings in that minute are dropped: they are neither written as active alerts nor delivered to On-call.
  • What is counted: new-alert-object attempts per rule per calendar minute, including the attempts that get dropped.
  • Active alert objects are unaffected: their repeat notifications and recovery are handled as before; the rule just stops creating new ones.
  • One summary alert on entry and one on release: when the limit trips and when it clears, the rule’s collaboration space gets a summary alert naming the rule, the limit and the start time. It follows the account language and is not repeated.
  • No backfill: dropped firings are not replayed after the limit clears, and the number dropped is not reported separately.
  • Release condition: after the last over-limit window, five complete minutes (five windows) without going over again release the rule automatically.
These summary alerts usually mean the rule’s alert dimensions include a field that differs from row to row (log body, timestamp, Trace ID), so every row becomes a new alert object. Narrow the label fields, or aggregate in the datasource before alerting; see Query result field mapping and VictoriaLogs.