Set thresholds for an alert condition

When you create a condition, you set thresholds that determine what will open a violation. This document explains what thresholds are and how to set them.

What is a threshold?

For a condition, thresholds are the settings that determine what opens a violation. Depending on a policy's incident preference, a violation may result in:

  • The creation of an incident.
  • Notifications being sent.

Examples of thresholds:

  • An application’s average web response time is greater than 5 seconds for 15 minutes.
  • An application’s error rate per minute hits 10% or higher at least once in an hour.
  • An application’s AJAX response time deviates a certain amount from its expected baseline behavior.

Besides a mandatory critical threshold level, you can also set thresholds for a less serious warning level, which doesn't generate an incident or send a notification.

View and set thresholds

Thresholds are set during the process of creating a condition:

Goal Instructions
Set thresholds for a new condition Set thresholds as part of the process of creating a condition.
View and update thresholds for existing conditions To view a condition’s thresholds: find that condition in the UI. To update thresholds, select a condition’s thresholds and make changes.

To learn more about specific alert condition types (like baseline and NRQL), see Types of conditions.

Details about other functionality and rules:

Use the Alerts API

You can list and edit conditions with the Alerts API.

Set optional warning level

You can set thresholds for two levels: critical (required) and warning (optional).

Threshold level Details
icon-alert-critical.png Critical (red) Required. When a violation occurs, it will send notifications depending on incident preference settings.
icon-alert-warning.png Warning (yellow) Optional. Doesn't generate notifications. Use a warning threshold when you want to monitor when a system behavior is concerning or noteworthy but not important enough to require a notification.
Loss of signal (NRQL only)

A loss of signal is a period of time when no data is received by New Relic. This could be the result of an entity or service going offline, an issue with an agent or collector, or networking problems in a data center or the internet. You can use loss of signal detection to create a new violation and notification if a signal stops and you expect a service to be stable. You may also want to use this to determine when an ephemeral service stops and set the action to close any open violations that exist for this condition.

To learn more about signal loss, gap filling, and how to request access to these features, see this announcement.

Set time intervals

Different condition types have different minimum time intervals. For example, some condition types have a minimum time interval of 5 minutes (for example, APM metric alert conditions), and others have a minimum time interval of 1 minute (for example, NRQL alert conditions).

Set URL for runbook instructions

For how to set this via the UI, see Runbook instructions.

For more help

If you need more help, check out these support and learning resources: