The Service Map for APM is here!

What are recovery thresholds?

Definition

Recovery thresholds are optional thresholds added to a monitor to indicate an additional condition to a monitor’s recovery from alert or warning states.

Behavior

When you set up a threshold metric monitor, you get alerted when a metric passes the alert threshold.

The recovery threshold adds a condition to the monitor’s recovery such that it only enters recovered state once it has passed the recovery threshold. If no recovery threshold is set, the monitor recovers whenever the alert conditions are no longer met.

The recovery threshold is satisfied with the recovery condition is met. The recovery condition varies based on the alert condition:

Alert condition Recovery condition
> alert threshold <= alert recovery threshold
>= alert threshold < alert recovery threshold
< alert threshold >= alert recovery threshold
<= alert threshold > alert recovery threshold

Use case

Recovery thresholds reduce noise of a flapping monitor. This can increase confidence that upon recovery, the alerting metric has recovered and the issue is resolved.

How to set up recovery thresholds?

Website UI

Set the alert or warning recovery threshold when creating a monitor under Set alert conditions:

Update this

API

When you create/edit a monitor via the API, use the attributes critical_recovery and warning_recovery in the options.thresholds attribute of your JSON monitor:

"thresholds": {
                "critical": 80,
                "warning": 60,
                "critical_recovery": 70,
                "warning_recovery": 50
              }

Further Reading

Additional helpful documentation, links, and articles: