A service level agreement (SLA) is the contract between you, as a provider, and your customers, that defines the expected level of service for your web applications. SLA usually refers to the expected performance and availability of your application.
A site may be required to have uptime of 99.95% of the year excluding maintenance and the response time within 2 seconds.
With AlertSite, you can monitor SLA compliance for your key web pages, transactions, and APIs. You can set SLA objectives for your monitors, and then see how the actual results compare to your SLA goals.
This topic describes the terms used in AlertSite SLAs. To learn how to configure SLAs, see Create a Service Level Agreement.
Service level objectives
An SLA in AlertSite includes service level objectives (SLOs) for availability, uptime, and response time.
The availability and uptime percentage goals can be defined with up to three decimal places, such as 99.995 %.
The response time goal can be defined with up to two decimal places, such as 1.45 seconds.
Performance (response time)
The SLA performance goal is the maximum acceptable response time for a monitor. Technically, the goal is set on the average time measured from multiple locations at the same time. For example, the SLA goal of 3 seconds means the average response time reported from all monitoring locations should be 3 seconds or less.
Say, you are monitoring from 3 locations, and they report the response time of 2.1, 2.5, and 3.2 seconds. The average time is 2.6 seconds which matches the SLA goal. Even though one measurement is higher than the SLA goal, the average meets the goal so the site is SLA-compliant.
For multi-step monitors like real-browser (DéjàClick) transactions, the SLA goal includes the total response time of all test steps.
For website and real-browser monitors with the fullpage monitoring enabled, you can use the basic response time or the full page response time as the SLA goal. The full page time includes the load time of both the base page and all images, CSS, scripts, and other page objects.
SLA performance goal is also used as the Apdex threshold for that monitor.
Availability is the percentage of successful tests out of all tests. A successful test means the monitored website is reachable, returns expected results, and does not experience errors. Successful tests have status 0. If a specific location experiences an error, the site is considered unavailable from that location. This way, availability reflects the regional perspective on webpage reliability.
Availability objective accounts for the allowed number of failed tests. For example, a 95% availability goal allows for 1 failed test out of 20 tests.
During each monitoring interval, a site can be up or down. A site is up if any monitoring location performs a successful test, even if other locations reported errors. A site is down if all locations report errors during the same monitoring interval. Uptime is the percentage of time your site is up in a given period, excluding periods not covered by the SLA (such as scheduled maintenance). Uptime is not the same as availability: a site can be up, but not available in some locations due to network outage.
Uptime percentage can be greater than or equal to availability.
|Note:||Uptime monitoring is available only for monitors with the SLA (MultiPOP) monitoring mode. It is not available for monitors that use private monitoring locations (Private Node Server or InSite).|
Difference between uptime and availability
Uptime means the site was tested successfully from at least one location. For example, if 4 out of 5 locations were unable to access the site during the same interval, availability is 20% but the site uptime is 100% – because 1 location could still access it.
Below is a more detailed example with two locations. In examples #2 and #3, one of the locations could not access the site. This is reflected in the availability percentage. The site is still shown as up, because another location could reach it at the same time. The site is down only one time, at #4 – when both locations report an error at the same time.
(4 out of 5 times)
SLA operating periods
SLA operating periods are the days of week and hours when the SLA is enforced. While many applications have a 24/7 operating period, some application SLAs are only enforced during the business hours, for example, Monday – Friday, 9 AM – 5 PM. You can also define one-time exclusions for scheduled maintenance and other periods when your performance and availability may be impacted. For more information, see SLA Operating Periods and Exclusions.
SLA compliance is calculated using the data from the SLA periods only. At other times monitoring still happens but results are not included in SLA calculations. If you want to disable monitoring during specific periods, you can use monitor blackouts.