Configure Alerts

Need to update the new checbox under Lambda Events "Do not send alert notifications for traces marked as warning"

With Epsagon, you can get notified for every problem the occurs in one of your functions, to any supported alert integration.

Epsagon can alert for any of the following metrics:

Metric

Threshold Alert

Singular Alert

Invocations

Supported

Errors

Supported

Supported

Timeouts

Supported

Supported

Out of Memory

Supported

Supported

Cost

Supported

Duration

Supported

Memory

Supported

Singular alert

Defining a singular alert means that you'll get a notification for every time it happens (with an aggregation up to 5 minutes). For example:

  • Time: 11 am, function X got an error - you'll get notified immediately, for 1 error occurrence.
  • Time: 11:02 am, function X got an error - aggregating.
  • Time: 11:04 am, function X got an error - aggregating.
  • Time: 11:05 am - you'll get notified, for 2 error occurrence.

The notification details include the number of occurrences.
To define a singular alert, got to the alerts screen, and create a new "simple" alert. In the alert type - choose the ones that you're interested in.

Threshold alert

A threshold alert is defined by the following parameters:

  • Event - the desired metric (example: duration).
  • Aggregation - the aggregation function to apply on the metric (example: average).
  • Timeframe - the period to conduct the aggregation on the metric (example: a day).
  • Operator - the operator which will be used to check the threshold (example: >=).
  • Value - the threshold value to cross (example: 300 ms).

Defining a threshold alert means that you'll get a notification for every time the threshold is crossed.
In the above example, we will get a notification when our function's average duration in a day >= 300 ms. When it returns to be < 300 ms, we will get another "OK" status notification.

To define a threshold alert, got to the alerts screen, and create a new "threshold" alert. fill in the parameters you're interested in.

Alert for a Group of Functions

It is more convenient to treat a group of functions in the same alert. /*It provides the ease of:

  • Aggregating all alerting logic into one element rather than several functions.
  • Alert auto applies to any new function that joins the application.

You can configure the application name through the tracing agent, or through the Lambda Dashboard.


Did this page help you?