Metrics User Guide

A metric is a quantifiable measurement whose values are monitored by Metricly and used to assess the performance of an element. Metrics are always associated with one or more elements.

Examples of metrics:

  • CPU Utilization
  • Network Bytes per Second
  • Response Time
Collected Metrics

Metricly has collected metrics for every element in the application. Policies can be created using collected metrics to determine if data is missing through alerts, and further configured to fire notifications. After 15 minutes of missing data, the element suspends and the alert stops. At this point, an external event is posted to the event timeline noting that Metricly has not received data.

Collected Metric Example
Fully Qualified Name(FQN) Description Units BASE CORR UTIL
netuitive.metrics.collected.percent % of metrics that collected data in the last interval.

Computation:
((# of collected metrics) / (total # of metrics) * 100)

Percent
Computed Metrics

To see the various collected and computed metrics Metricly monitors, you’ll have to look at the metrics section of an integration page.

Metric Page Features


Filters

Contains several filters where you can search for element names, metric names, tags, attributes, and more.

  • Name Contains: Specific search based on name
  • Search Metrics: Broad filter based on metric type, provided in tree and list form.
  • Element, Type, More:  Additional expandable filter options.

Once you have added all needed search criteria, click Render Charts to populate results.


Group By

Contains a menu of all quick groups, tags, and attributes available to Metricly. After you’ve rendered some metric charts, choose one of the groupings and Metricly will automatically group the metric charts based on your selection. If you want to remove a selected grouping, click the x next to the grouping.

If you chose to display all EC2 metrics and then grouped by instance ID, each instance’s set of metrics would be grouped in a section together under the instance’s ID.


Quick Filters / Metric Charts

Click a Quick Filter to render metric charts automatically for the selected filter. Metric charts are a visualization of a metric for an element. See below to learn more.


Events Graph

The Events graph displays the events in your environment based on the Time Frame setting and other search filters. For more information, see Events graph. Click an event to have the option of viewing the violating metrics on the Metrics page or edit the policy the event is associated with.


Metric Chart Size

Select how wide or how tall you want your metric charts to be.


Time Frame Controls

The Time Frame controls the range of data displayed. To refresh data, click the refresh button. Selecting “1w” in the Time Frame displays the most recent week of data and/or elements. By selecting “Ending Now,” you can specify a range of data beginning with a date other than today. For more information, see Time Frame.


Data Aggregation

Select the type of data displayed on the metric charts.

Option Description
Raw Shows data that has not been aggregated by Metricly.

Computed metrics do not have raw data because they are calculated by Metricly.
5 Min Displays aggregate data that Metricly generates by averaging the data collected from a given integration at 5 minute intervals.
1 Hour Displays aggregate data that Metricly generates by averaging the data collected from a given integration at 1 hour intervals.

Chart Options

Show and hide baseline bands, contextual bands, and data points; normalize the Y-axis per metric; or display step charts. Details on data points below.

DP Color Description
Black Black data points indicate aggregate data or data that has been interpreted by Metricly. To generate aggregate data, Metricly averages the data collected from a given integration at 5 minute intervals.
Gray Gray data points also indicate aggregate data but for data collected that Metricly averages at 1 hour intervals.
Green Green data points indicate sparse data or data generated by Metricly when data is not collected from a integration within a given period of time. The metric value for sparse data is always 0.
Red Red data points indicate a deviation or data that falls outside the learned Contextual or Baseline bands.
Blue Blue data points indicate raw data or data that has not been interpreted by Metricly. Because raw data is not interpreted by Metricly, it does not produce deviations.

Metric Charts

Metric charts on the Metrics page allow you to view time-series data collected on elements in your environment. The range of data shown in metric charts corresponds to the range of data shown in the Events graph. This allows you to compare event data to the metric behavior that caused it.


Metric Metadata

Click the metric name to open the metric metadata panel.


Element

Click the element name to open that element’s Element Detail panel in the Inventory Explorer.


Sub-menu

Offers the following options:

  • Metric Unit: Change the displayed unit (UI only; does not affect default unit metadata).
  • Aggregation(s): Select which aggregate values you want to display on the metric chart. Avg is the default. Each aggregation corresponds to a different color. Computed metrics do not have aggregations available.
  • Download Buttons: Download the current chart for this metric to a .png.jpg, or .svg file.
  • Add to Dashboard: Add any metric chart to the desired dashboard as a single metric time series widget (see below).
  • Other Elements with this Metric: View other elements with the same metric and open the respective elements’ metric chart (see below).

Add a Metric Chart to a Dashboard (Single Metric Widget)
  1. Open the Metric chart sub-menu.
  2. Click Add to Dashboard.
  3. In the Add to Dashboard window, click Dashboards and select a dashboard or type to search.
  4. Change the widget name if desired.
  5. Click Save.

The widget is now available on the selected dashboard.


Add a Metric Chart with Multiple Metrics to a Dashboard (Multi-Metric Widget)
  1. Merge as many metrics as desired.
  2. Open the Metric Chart sub-menu (. . .).
  3. Click Add to Dashboard.
  4. In the Add to Dashboard window, click Dashboards and select a dashboard or type search.
  5. Change the widget name if desired.
  6. Click Save.

The widget is now available on the selected dashboard


View Charts of the Selected Metric Type Collected on Other Elements
  1. Open the Metric Chart sub-menu (. . .).
  2. Click Other Elements with this Metric.
  3. Select as many elements as desired.
    • Click All to select all the elements with the selected metric.
    • Click None to clear all elements with the selected metrics.
    • You can also search for an element using the Search Elements field.
  4. Click Done. The metric charts open in the Metrics page.

Metric Metadata Panel

The metric metadata panel provides valuable information about the metadata associated with your metrics.


Metric Value Summary

The aggregate values of the last five-minute analysis cycle. (Metricly aggregates all of a metric’s value every five minutes.)

Summary Description
Min. The minimum value recorded in the last cycle.
Avg The average value recorded in the last cycle.
Max. The maximum value recorded in the last cycle.
Count The amount of values recorded in the last cycle.
Sum The sum of values recorded in the last cycle.
Value Used The type of aggregation used in the analysis cycle
The time frame setting does not affect the values.

Updated

The time the last five-minute cycle occurred in which the data was updated.


FQN

The fully qualified name (FQN) of the metric.


Id

The full ID of the metric.


Metadata

Lists metadata associated with the selected metric.

Metadata Description
Units The type of measurement the metric is collected in.
Type The type of metric; currently, the only two values are “collected” or “computed.”
Valid Min If set, the valid minimum value the metric can have.
Valid Max If set, the valid maximum value the metric can have.
Created When the metric was created.
Sparse Mode Defines the strategy for replacing missing data. Either missing data is replaced with a zero (ReplaceWithZero) or no strategy is taken at all (None).
Baselined If the Baseline Band analysis is enabled or disabled for the metric.
Correlated If the Contextual Band analysis is enabled or disabled for the metric.
Historical Min Historically, the lowest value the metric has experienced.
Historical Max Historically, the highest value the metric has experienced.

Description

The friendly name of the metric.


Go to

Allows you to navigate to the metric on the Metrics page (if you’re in the Element Detail panel) or to the Element Detail panel of the associated element (if you’re on the Metrics page).

Moving and Manipulating Metric Charts

Metric charts offer functionality that allow you to merge two charts together, move charts around on the Metrics page dashboard, zoom in, and drop a marker on a data point.

Merge Metric Charts
  • Click on the header of one metric chart and drag it on top of another chart into the green “Merge” region.
    click to unmerge the metric charts.
Move Metric Charts
  • Click on the header of one metric chart and drag it into the blue “Move” region located above a different chart.
    The chart will take the place of the chart to which it is dragged.

Zoom in on a Chart
  • Click and drag across the area on a chart that you want to view.
    When you zoom in on one chart, Performance Explorer will automatically zoom in on all charts in the same location.

Drop a Marker
  • Click on a chart in an area that aligns with a data point.
    When you drop a marker on one chart, Performance Explorer will automatically drop markers on all charts in the same location. Note that you will only be able to drop a marker on a chart if you’re clicking outside of the Baseline and Contextual bands.

Computed metrics

Computed metrics are metrics that Metricly calculates using the data from multiple metrics collected on a integration. Computed metrics are used to compile utilization reports of your environment and to specify metric behavior in default policies. A metric can be identified as a computed metric if its name begins with “metricly.”

For example, a CPU utilization metric belonging to an EC2 element collected from an AWS integration is named aws.ec2.cpuutilization, while an average latency metric computed by Metricly for EBS volumes is named netuitive.aws.ebs.averagelatency.

Troubleshooting

Collected Metrics

If you are seeing metrics that will not collect continually, there are two common scenarios: your agent-based (Linux, Windows) collector has metrics that are not available anymore being collected or your agentless-based collector (AWS, Azure) is collecting metrics from a service that’s not being used, e.g., a lambda function not firing or an SQS queue is empty.

Removing these metrics from being collected is the optimal solution. You can use a metric whitelist or blacklist for agent-based collectors; for agentless-collectors, you may need to tweak your filtering in Metricly or you may be experiencing ingestion issues.

An example blacklist: you could use a negative lookahead (?! ) to specify a group that cannot match after the main expression–if something matches, the result is discarded: metrics_blacklist = elasticsearch.indices.(?!_all$|datastore$|docs$)