Alerts: Configuration and Best Practices

Table of Contents

• Types of Alerts
• Rules Based Alerts
  • Create an Alert
  • Edit Alert
  • Set Maintenance Window
  • Bulk Actions
  • Alerts Log
• Synthetic Error State Tracking
• Workflow
  • Tuning and Maintaining Alerts

Types of Alerts

Standard Rules Based Alerts

These alerts are mainly performance alerts. With these alerts you can do things like monitor onload on a home page. For these alerts, you can set thresholds for when the alert should fire, such as a percent change, standard deviation or fixed value threshold. Rules-based alerts can be found in the main Alerts page.

Synth Error State Tracking

This type of alert is tied to synthetic monitors, and can be configured in the monitor settings. You can configure the lookback window, how many different locations you are required to trigger an alert, etc. 

SLA Alerts

Found in the Tag and Content Governance module, these are alerts for first and third party content on your site. For example, you could monitor how Google Analytics or Blue Triangle is affecting your site. You can also monitor things such as file size or average duration, and you can alert on when those tags are exceeding the set thresholds.

Missing Services Alert

Alerts on services or tags that fall off the site or are not loading as often as they should be loading on your site.

Other Notifications

These can all be configured under your user profile. If you click edit profile and scroll down you’ll see check boxes, all of these options are different forms of alerting.

Rules-Based Alerts

Alerts Page

The Alerts page can be found in the left hand navigation menu, under Alerts and Automated Reports.

Create Alert

There are four steps in the alert creation process, follow the interactive demo bellow for a step by step guide on creating alerts in the portal. To start, click button to create alert and the wizard will pop up to help.

First are the general settings. You must give the monitor a unique name, the select recipients. The alert can be delivered to an email, notification group or SMS number. There is also the option to create a new notification group, integrate with Slack, or other integrations such as PagerDuty. You have the option to embed a graph in the alert email itself, where you can see at a glance what is going on before investigating. Customize the email subject and body of the email, or leave blank for default. 

Next set the conditions for the alert, such as data type (RUM, Synthetic or API) and thresholds, which will differ by data type. Thresholds can be set to a fixed value, percent change, or standard deviation. Percent change allows you set a positive and negative percent change. Violation count thresholds are for synthetic and API alerts only. Set the  metric to alert for, and select to notify when alert clears or not. Set warning and critical thresholds, may need to be adjusted.

Schedule it, enable a certain time range or run indefinitely, set time zone, enable certain days od the week, enable reminders, maintenance windows. Set evaluation window, amount of time for the alert to look back to establish baseline, and minimum sample size, number of samples required during evaluation window. For example, in the last five minutes there must have been 100 page views before it can alert.

Finally, select any filter options such as traffic segment, page name, browser or device type, location, etc.

Select Finish and the alert will either start running immediately or at the scheduled time, if that is the case

Edit Alert

To edit an alert at any time, click the pencil icon next to the alert name in the Alerts page.

Set Maintenance Window

Maintenance windows give you the ability to silence alerts during scheduled maintenance or certain periods where maybe traffic. Periods of low traffic can create false positives for alerts. Follow the interactive demo below for step by step instructions on creating maintenance windows. 

Bulk Actions

You can apply a maintenance window, toggle reminders, enable, disable, delete, pause alert immediately, or schedule a pause to more than one alert at a time using bulk actions. The difference between pausing an alert and applying a maintenance window is that the pause is a one time event while a maintenance window is a reoccurring daily or weekly time that alert is silenced.

Alerts Log

The Alerts Log is not only a log of the state changes for each alert, but also a log of any edits or updates to the alerts themselves. This allows for more context on situations like why an alert cleared, or if you need to check for edits like threshold changes

Synthetic Error State Tracking

The Error State Tracking Page shows you all the errors that have occurred on your synthetic monitors, as well as other related information, such as the duration of each error. Follow the interactive demo below for step by step instructions on configuring Error State Tracking for a monitor. 

While Alerts can be made for all data types, Error State Tracking can only be found in the configuration of synthetic monitors. 

Workflow

Tuning and Maintaining Alerts

Every Alert is actionable, whether it be an issue or non-issue. Of course, if there is a real issue on the site, you may go ahead and investigate the issue. However, if you are receiving an alert that is a non-issue to you, you may want to follow up by either removing yourself from the alert distribution list if the alert does not apply to you, fine tuning the alert, or creating a maintenance window. Sometimes alerts might require changes to it’s warning, critical, and clear thresholds to better support the needs of the site and your team.