WAN Insights and ThousandEyes Alerts
Last updated
Last updated
Due to recent platform-wide naming, navigation, and URL changes in the product, you may notice some discrepancies between the product and the screenshots displayed in our technical documentation. The instructions and actual pages in the product are still valid and haven’t changed. Please bear with us as we update our screenshots to better match the in-product experience. See the full scope of changes on .
The ThousandEyes feature generates alerts and notifications when specified error or network conditions are present, based on alert rules that you create. When an alert rule is triggered, a list item displays in the Alert List page in ThousandEyes. For example, you can set WAN Insights alerts to call attention to changes in path quality or capacity utilization.
WAN Insights alert rule types are Capacity and Site Quality. Within each alert rule, you can add one or more conditions, which can be combined together.
If you need alert conditions to trigger separately, configure two separate alert rules of the same type.
Alert rule types are either capacity or site quality. You can’t have a single alert rule that focuses on both.
Within each alert rule, you can use alert conditions to further focus on specific subsets of sites, numbers of users, or specific applications to focus on.
To configure alert rules for WAN Insights, go to Manage > Alert Rules and select the WAN Insights tab. Click Add New Alert Rule. You can find documentation for ThousandEyes alerts in the area of the ThousandEyes user interface.
If you want visibility into WAN Insights alerts outside of the ThousandEyes platform itself, you can use custom webhooks to send notifications to other systems as described in . You can also add alerts to custom and embed those dashboards in other systems.
For WAN Insights, the alert types are:
Site Quality. Shown in . It’s the path quality described on .
Capacity. Capacity utilization as described on this page. Data is broken out in more granularity on the following screen,
Use a site quality alert to identify sites experiencing poor path quality.
For the Site Quality alert type, the available alert rule conditions are:
Number of Users. Minimum/maximum number of active users on a site in order for it to be considered.
Site ID. Use to restrict alerting to a particular site or set of sites.
Use a capacity alert to call out capacity saturation affecting particular hostnames and/or interfaces.
For the Capacity alert type, the available alert rule conditions are:
Hostname. Use to focus on capacity conditions involving a particular hostname.
Interfaces. Choose one or more interfaces from the drop-down list with check boxes. For example, if you are using interfaces biz-internet and gold, you could choose to focus this alert rule only on capacity saturation for biz-internet.
What does “x out of n times in a row” mean for WAN Insights alerts? It’s based on a WAN Insights alert rule evaluation that occurs in a particular round.
The first number x is the minimum number of occurrences
The second number n is the overall amount of rounds you are considering.
The term “round” here refers to different time intervals for each WAN Insights alert type, as follows:
For capacity data, where the Alert Type is specified as Capacity | Utilization, note that a WAN Insights “round” is equal to 24 hours.
For path quality data, where the Alert Type is specified as Quality, note that WAN Insights calculates quality on an hourly basis – i.e., a single round for quality data is equal to 60 minutes. Therefore, by selecting 3 out of 5 times in a row, will mean that an alert will only fire if the conditions are met in 3 rounds, at minimum. Thus, the soonest an alert will fire will be every 3 hours. If you want an alert to fire sooner, adjust the first number accordingly.
WAN Insights alerts are evaluated on an hourly or daily basis, depending on the alert type. ThousandEyes compares the alert rules that you create with the data collected from WAN Insights. If this comparison shows that a condition meets or exceeds an alert rule threshold, then the alert will trigger.
As a reminder, capacity rounds are daily and quality rounds are hourly.
An alert will be cleared whenever data no longer matches the corresponding alert rule, or after 36 hours in the case of no-data.
One of the conditions for the WAN Insights alert type of Capacity is a condition called Percentile that’s designed to exclude outliers. This percentile is not a percentage, and it’s not related to the path quality measure.
Note that the percentile condition also includes the actual capacity saturation percentage as the second value. You should always have a percentile defined for this alert type, as one of the alert rule conditions.
The alert rule for capacity lets you specify a different percentile than the one used on the Capacity Planning screen on the Settings tab, and the alert rule is based on the past 30 days. In addition to the percentile, the alert rule condition also includes a percentage which relates to the percentage of capacity used.
For example, we could create an alert rule for capacity that specified the following:
95th percentile that is greater than 80 %
In this example, the alert will trigger when the 95th percentile of utilization is greater than 80% of capacity over the past 30 days.
A few ways you can use WAN Insights capacity alerts:
Specifying an alert rule with a higher percentile takes more extreme situations into account. For example, the 100th percentile would consider every data point including outliers, whereas the 95th percentile would take the highest values excluding the topmost 5%.
Specifying a different alert rule with a lower percentile, say 95 instead of 98, will alert for degradation that is more pervasive and frequent.
You might want to have alerts for different percentiles as an advance warning.
Contractually you might have an SLA for 95th percentiles or 98th percentiles.
Percentage. Site quality itself, expressed as a percentage. See .
App Class. See . Application categories or classes are bundles of applications with similar traffic characteristics.
Percentile. Use to exclude outliers from your capacity alerts. Outliers represent conditions that may occasionally cause saturation events, but are not indicative of the most frequent or prevailing patterns of capacity utilization. See further down on this page.
WAN Insights data comes from your SD-WAN, as described in .
Path quality data, as referenced in the alert rules for path quality, come from synthetic network probes as described on .
Capacity data has two parts. Bandwidth settings are ingested, and usage data subsequently collected, as described in .
The percentile relates to the main . On the capacity planning main screen, in order to eliminate outliers, we use percentiles also known as utilization aggregates, which are explained on the Measurement and Aggregation section of the WAN Insights page.
