Open topic with navigation
You can add service level agreements (SLA) to your monitors to define the acceptable parameters of operation for your websites and APIs. You can set SLA objectives for uptime, availability and response time (or full page response time). Once set, you can use AlertSite dashboards and reports to review the monitor metrics as compared to the SLA goals.
Note: You need to be an Admin or Co-Admin user to create or edit SLAs.
To monitor SLA compliance, you first need to create monitors for your key pages and transactions.
If you want to monitor uptime in addition to availability, configure the monitor with the SLA (MultiPOP) monitoring mode and at least 2 locations. This mode is available for all website, real-browser and API monitors except those using private monitoring locations (Private Node Server or InSite). You can choose to rotate locations, with 2 or more locations per interval.
Uptime monitoring requires multiple locations, because locations coordinate with each other to see if the site is up or down globally. For example, if two out of three locations can reach your site, it is still considered up. It is only considered down when no locations can reach it.
You can also use other monitoring modes – Global Verify, Primary and others. In this case you will be able to set SLA for availability and response time, but not uptime.
Open the list of SLAs in the top left, and click New SLA.
Select the monitor for the new SLA, and click Next.
On the next screen, enter the SLA name, description and the SLA objectives:
Availability Goal – The minimum acceptable percentage of successful tests. For example, 95%.
Uptime Goal – Only for monitors with the SLA (MultiPOP) monitoring mode and at least 2 locations. Specifies the minimum acceptable uptime, as percentage. For example, 99%. A website is “up” if at least one location performed a successful test during the monitoring interval. Uptime goal can be equal to or greater than the availability goal.
Performance Goal – The maximum acceptable response time, in seconds. Also used as the Apdex T threshold between satisfying and tolerating performance. For multi-step monitors, this includes the total time of all steps.
Secondary Apdex Threshold – Optional. The Apdex F threshold between the tolerating and frustrating performance. Must be greater than the Performance Goal. If left blank, defaults to four times the Performance Goal. For multi-step monitors, this value includes the total time of all steps.
Use fullpage response time for performance goal (when available) – Only for website, real browser (DéjàClick) and mobile web monitors with fullpage monitoring enabled. Select this option to apply the SLA performance goal to the fullpage response time (the load time of the entire page with all embedded objects, images, scripts and stylesheets). If not selected, the SLA is applied to the base page response time.
AlertSite UXM will calculate SLA compliance and show it on the dashboard.
SLA operating periods are the days and hours when the SLA is in effect. The default operating period for new SLAs is 24/7. However, some application SLAs are only enforced on specific days and hours, such as business hours (Monday–Friday, 9 AM–5 PM). If so, you need to configure your SLA schedule appropriately. Data falling outside of SLA operating periods will not be included in your SLA calculations.
To define SLA operating periods:
For each day, select the hours of the day covered by the SLA.
To exclude specific days like weekends, select the Exclude check box next to them.
Below is an example for Monday–Friday, 9 AM–5 PM.
SLA exclusions are one-time instances when the SLA is not in effect, such as during scheduled maintenance. Monitoring stills happens, but the results are not included in your SLA calculations.
To add an exclusion:
Specify the start time, end time and a description. You can add both future and past dates.
Monitoring data from these periods will now be excluded from SLA calculations.
Monitors with an SLA follow the usual AlertSite alerting rules. For example, they send availability alerts on HTTP errors, missing keywords and other conditions specified in the monitor settings. To receive response time alerts, you will need to set performance thresholds in the monitor settings as explained in Performance Alerts.
Alerts are always based on the test results, rather than on how the results compare to the SLA goals. This is because the SLA goal is an average over a given reporting period – day, week, month and so on. Typically, a spike in the response time can usually be ignored as long as the average is OK.
Monitors with the SLA (MultiPOP) monitoring mode have an extra option to control availability alerts – you can specify the minimum number of error locations for an alert to trigger. You can choose to be notified if 1, 2, 3 or all monitoring locations detect an error at the during the same run interval.
Choose when availability alerts should be sent:
For example, to get an alert only when ALL locations see an error, select Suppress notification if ANY location can reach your site. In this case if only one location sees an error you will not be notified.