Introduction

This document provides step-by-step instructions on how to configure alert thresholds for SaaS application metrics. Alerts are triggered based on the thresholds set by users, allowing for proactive monitoring and issue detection.

Supported Operators for Alerting

The following operators can be used for setting alert thresholds:

  • == (Equal)
  • >= (Greater Than or Equal)
  • > (Greater Than)
  • < (Less Than)
  • <= (Less Than or Equal)
  • != (Not Equal)
  • In Range
  • Out of Range
  • In List
  • Not in List

Configure Alert Thresholds

  1. Equal (==)
  • A Warning Alert is triggered when the instance value equals the Warning Threshold.
  • A Critical Alert is triggered when the instance value equals the Critical Threshold.

Example:

If possible metric values are 0,1, and a Critical Alert is required at 0, set the Critical Threshold to 0. If only a Critical Alert is needed, set the Warning Threshold to 0 or 0.5. Similarly, to trigger a Warning Alert at 1, set the Warning Threshold to 1.

  1. Greater Than or Equal (>=)
  • A Warning Alert is triggered when the instance value is greater than or equal to the Warning Threshold.
  • A Critical Alert is triggered when the instance value is greater than or equal to the Critical Threshold.

Example:

If a Critical Alert is required at 40 or higher, set the Critical Threshold to 40. To trigger a Warning Alert at 30 or higher, set the Warning Threshold to 30.

  1. Greater Than (>)
  • A Warning Alert is triggered when the instance value exceeds the Warning Threshold.
  • A Critical Alert is triggered when the instance value exceeds the Critical Threshold.

Example:

If a Critical Alert is needed when the instance value exceeds 70, set the Critical Threshold to 70. Similarly, to trigger a Warning Alert above 55, set the Warning Threshold to 55.

  1. Less Than (<)
  • A Warning Alert is triggered when the instance value is below the Warning Threshold.
  • A Critical Alert is triggered when the instance value is below the Critical Threshold.

Example:

To trigger a Critical Alert below 80, set the Critical Threshold to 80. Similarly, for a Warning Alert below 130, set the Warning Threshold to 130.

  1. Less Than or Equal (<=)
  • A Warning Alert is triggered when the instance value is less than or equal to the Warning Threshold.
  • A Critical Alert is triggered when the instance value is less than or equal to the Critical Threshold.

Example:

To trigger a Critical Alert at 60 or lower, set the Critical Threshold to 60. Similarly, for a Warning Alert at 85 or lower, set the Warning Threshold to 85.

  1. Not Equal (!=)
  • A Warning Alert is triggered when the instance value does not match the Warning Threshold.
  • A Critical Alert is triggered when the instance value does not match the Critical Threshold.

Example:

If possible metric values are 0,1,2, and a Critical Alert is required at 0, set the Critical Threshold to 1. Similarly, to trigger a Warning Alert at 2, set the Warning Threshold to 1.

  1. In Range

Accepts two numeric values separated by a comma (,).

  • A Warning Alert is triggered when the instance value is within the Warning Threshold range.
  • A Critical Alert is triggered when the instance value is within the Critical Threshold range.

Example:

For metric values 0-6, if a Critical Alert is required at 1, set the Critical Threshold to 0,2. To trigger a Warning Alert for 2-5, set the Warning Threshold to 1,6.

  1. Out of Range

Accepts two numeric values separated by a comma (,).

  • A Warning Alert is triggered when the instance value is outside the Warning Threshold range.
  • A Critical Alert is triggered when the instance value is outside the Critical Threshold range.

Example:

For metric values 0-6, if a Critical Alert is needed for 0,4, set the Critical Threshold to 1,3. Similarly, to trigger a Warning Alert at 1,7, set the Warning Threshold to 2,6.

  1. In List

Accepts multiple values separated by a comma (,).

  • A Warning Alert is triggered when the instance value matches any of the Warning Threshold values.
  • A Critical Alert is triggered when the instance value matches any of the Critical Threshold values.

Example:

For metric values 1-6, to trigger a Critical Alert for 1,3,5, set the Critical Threshold to 1,3,5. Similarly, to trigger a Warning Alert at 2,6, set the Warning Threshold to 2,6.

  1. Not in List

Accepts multiple values separated by a comma (,).

  • A Warning Alert is triggered when the instance value does not match any of the Warning Threshold values.
  • A Critical Alert is triggered when the instance value does not match any of the Critical Threshold values.

Example:

For metric values 0-4, if a Critical Alert is required at 0,3, set the Critical Threshold to 1,2,4. Similarly, for a Warning Alert at 4, set the Warning Threshold to 0,1,2,3.

Supported Placeholders in Alerts

When configuring a monitor, you can customize the Alert Subject and Alert Description fields. To make alerts more meaningful and dynamic, you can use predefined placeholders that automatically insert values such as resource details, metric information, or event data.

By default, the Subject and Description fields are prepopulated with commonly used values. If needed, you can further customize these fields by using supported placeholders.

These placeholders are replaced with actual values at runtime when an alert is triggered.

PlaceholderDescription
${resource.name}Inserts the name of the resource associated with the alert.
${resource.type}Inserts the resource type.
${metric.name}Inserts the name of the metric that triggered the alert.
${metric.value}Inserts the actual metric value at the time of breach.
${threshold.value}Inserts the threshold value that was breached.
${alert.severity}Inserts the severity of the alert.
${alert.timestamp}Inserts the timestamp when the alert was generated.
${integration.source}Inserts the source system or integration that generated the alert.

Polling and Repeat Count

In metric monitoring, two key parameters control how data is collected and when alerts are triggered:

  • Polling Interval (mins)

    • Definition: The frequency (in minutes) at which the monitoring system collects metric data from a device, application, or service.
    • Purpose: Determines how fresh and up-to-date the monitoring data is. Shorter polling intervals provide more granular visibility but may increase load on the monitored system and the monitoring platform.
  • Repeat Count

    • Definition: The number of consecutive times a metric must breach its configured threshold (during each polling cycle) before an alert is generated.
    • Purpose: Prevents alerts from being triggered due to temporary spikes or transient conditions. Ensures that only sustained issues raise alerts.

How Polling and Repeat Work Together

The Polling Interval defines when the system checks the metric, and the Repeat Count defines how many times in a row a breach must occur before generating an alert.

ScenarioPolling (mins)RepeatDescription
A101If a breach is detected at 12:00 PM, an alert is generated immediately at 12:00 PM.
B102If breaches occur at 12:00 PM and again at 12:10 PM, the alert is generated at 12:10 PM.
C152If breaches occur at 12:00 PM and again at 12:15 PM, the alert is generated at 12:15 PM.
D153If breaches occur at 12:00 PM, 12:15 PM, and 12:30 PM, the alert is generated at 12:30 PM.

Key Points

  • If alerting is enabled without specifying thresholds, the thresholds default to 0, which may cause false alerts.
  • If both Warning and Critical thresholds are set to the same value, only Critical Alerts will be generated.
  • For metrics with multiple components:
    • If component-specific thresholds are not set, an alert will be triggered if any component exceeds the specified threshold.
    • If component-specific thresholds are set, an alert will be triggered only when a specific component exceeds its designated threshold.
  • If no component filters are applied, monitoring data will be collected for all metric components. When filters are applied, monitoring data will be collected only for matching components.