To aggregate is to combine by adding a collection of data points and treat them as a single entity.
An alert is a notification of the occurrence in metrics of an anomaly, no data, or variance from a static value.
The difference between the peak metric value in an anomaly from the normal range.
The anomalies page.
An anomaly is a set of metric values that deviate from the usual behavior of the metric.
In alert notifications, a channel identifies a distribution mode, such as email, URL (Webhook), Slack or system user accounts. Each recipient within a channel is called an instance.
A composite metric is a metric that is the result of applying analytic function(s) on existing metrics. Once defined, Anodot treats composite metrics like other metrics in the system: learning normal behavior, discovering anomalies, and triggering alerts.
To correlate is to identify a mutual relationship or connection between metrics.
A dashboard is page in a user interface which holds the collection of current information of most interest to a user.
- A filtering mechanism to measure the difference between the peak value of the metric during an anomaly from the normal behavior, in percentage or absolute
- A function for metric expression. See Function Element Types.
To analyze and determine if an anomaly exists within a given set of metrics.
To detect an anomaly is to compare each new data point with the aggregated norm of prior points and to identify if the new data point is outside the normal limits for the metric.
Create a copy of an alert or dashboard for customizing.
Collection of values for metric properties and filters, possibly contained within a function, that identifies a target set of metrics.
Dialog that helps you dynamically build metric expressions. Using Anodot’s search, you are prompted to build an expression tree comprised of one or more expression nodes. Each expression node consists of property key-value pairs used to search for metrics and filters. Composite metrics are created when a function is wrapped around an expression or expressions.
Instance in Alert Channel
In alert notifications, the occurrence of a recipient within a channel.
A metric is the measurement of a characteristic that can be quantified over time.
The percent of change from the normal range to the peak value of an anomaly.
Rate at which a dashboard is refreshed with the latest data, in minutes or hours.
User-defined filter created and used in a Dashboard.
User-defined level for identifying the general severity of an alert. Severity is ranked as Critical, High, Medium, Low, or Info.
Anodot’s assessment of how important an anomaly is. Importance is determined by how much and how long the metric deviates from the expected pattern compared to the past anomalies of the metric. Significance is defined on a scale from 0 to 100, with 100 being the most significant.
The normalized upper and lower bounds of a metric. A sleeve is the background representation of normalized upper and lower values against which current observations are reported.
Measurement of how much variation there is within a set of metrics.
An arrangement to receive notifications for a specific alert.
The date range controls the time period of metrics or anomalies that are viewed in the dashboards, metrics and anomalies tabs. The range is according to the Last n Minutes, Hours, Days, Week, Months, and Year. You may also define a custom date range.
The span of time for which metrics are aggregated. Anodot aggregates metrics into 1 minute, 5 minutes, 1 hour, and 1 day intervals. Anodot also computes anomalies on each time scale independently. When viewing metrics in dashboards or the metrics tab, Anodot automatically selects the viewing time scale based on the Time Range selected. The time scale is marked with Auto. This setting can be overwritten.