If you choose to use the new UI when adding or editing conditions in an alerting policy, you reach the beta Select condition type page.
Near the top of this page is a bar that lets you revert to the current UI. See Creating an Alerting Policy for more information.
The Select condition type page lists the following choices:
- Metric conditions: Threshold, rate change, and absence conditions
- Basic health: Uptime-check health
- Advanced health: Process health
The interface described here applies only to the Metric Conditions type. The rest of this page assumes you have chosen that type. The Metric Conditions type covers most of the condition types offered by the current UI.
The page consists of two regions, a Target with an accompanying chart, and a Configuration region. If you are editing an existing condition, the page is populated. If you are adding a new condition, the page is empty.
The target is used to describe a metric in a condition you want the policy to monitor. The accompanying chart shows data from the time series in the metric, which can be helpful as you refine the target metric. The following screenshot shows a simple target metric and its chart:
The Target region uses the same metric selector that is used in Metrics Explorer and for creating charts. This metric selector gives you fine-grained control in creating the metrics for your alerting condition:
- You have control over aggregation and aligner choices, including with custom metrics and logs-based metrics.
- You can precisely filter and group the time series in the metric, including metadata filtering.
- You can edit conditions created with the API that cannot be edited with the current UI
The chart accompanying the metric selector gives you a visual representation of the data being captured by the metric as you configure it. If you are not already familiar with the metric selector, see Configuring a Target Metric.
In the current UI, various types of conditions are configured through separate condition options. With this metric selector, you can create them from a single UI. You can choose a metric for simple threshold based on, for example, the mean of the times series. A rate-of-change condition is similar, but based on a percentile aggregation.
For additional reference, Sample Policies includes a variety of alerting policies in JSON format.
The Configuration region describes the behavior of the target metric that will be considered as meeting the policy condition. Use this to specify how many of time series in the target metric (some, all, a percentage) must have some relationship (above, below, etc.) to a threshold value for a time period:
This configuration lets you specify whether the target metric will participate in a metric-threshold or metric-absence condition.
When you have completed describing your condition, click the Save Condition button at the bottom of the page. This takes you back to the Add (or Edit) Alerting Policy page, from which you can: