Monitor and Filter Page

The monitor and filter detail pages have information about the history of a metric as well as allowing the user to add alerts for the metric.

Navigating to the monitor and filter details

Once inside of a dataset pipeline, click on 1. the monitor listed in the monitors section in order to see the monitor details. Similarly, for datapoint pipelines click on the filter listed in the filters section in order to see the filter details.

30803080

Monitor and Filters details overview

30803080

Monitor details page with metric and alert history

30803080

Filter details page with metric and alert history

1. History

The history graph shows either the monitor metric history or a sample of the historical filter datapoints for a given partition and time range.

Monitors

The monitor metric history is composed of simple boxplots per time point where each boxplot shows the mean (centered dot), +/- one standard deviation (rounded rectangle), and the extremes (min/max whiskers). For timepoints where there is only a single sample in the aggregate, an outlined circle with a center dot is shown at that value.

Filters

The sample of filter historical datapoints shows the value for those datapoints in addition to the allowed bounds for the filter. Points that are considered passing are colored in green and points that are considered failing are colored in red. The overall behavior of the filter is displayed in the bottom as a series of bars showing the total number of points passing through the filter and the number of points considered anomalies (i.e. failing).

2. Alerts

The alerts section show the monitor metric history or filter metric (e.g. passing percentage) along with the allowed values for that metric per alert. New alerts can be added by clicking on the New Alert button in the upper right corner of the section. Each alert graph is shown as a separate item in the list. The alert graphs are similar to the monitor metric history graphs. They are composed of simple boxplots (see History above). In addition, at the bottom of each alert graph, there is a summary of the total number of events and the number of alarms at that point in time.

3. Monitor and Filter settings

The monitor or filter configuration can be seen by clicking on the information icon. Similarly, the list of available actions (e.g. Delete) is shown in the actions menu when clicking on the three dot icon.

4. Partition control

The partition control allows you to select the current partition that is being viewed in both the history and alert graphs. For dataset and datapoint pipelines that have been partitioned, each unique partition that has been seen is listed as a row in the table where each cell corresponds to a value for a given feature. Simply select a different partition, and the graphs will update to show you the data specific to that partition. For pipelines that have not been partitioned, the control is disabled.

5. Time range control

The time range control allows you to select the current time range that is being viewed on the x-axis of both the history and alert graphs. You can select between seeing the most recent data for the past minute, hour, day, or month, in addition to selecting a custom date range.

6. Alert settings

The configuration for each individual alert can be seen by clicking on the information icon on the right side of the alert. Similarly, the list of available actions (e.g. Delete) is shown in the actions menu when clicking on the three dot icon.