Action: Monitor Service Level Agreement (SLA)
Note:
Dependency requirement: A policy that includes this action must also include the Identify Consumer action.
This action is similar to the Monitor Service Performance action. Both actions can monitor the same set of run-time performance conditions for a virtual service, and then send alerts when the performance conditions are violated. This action is different because it enables you to monitor run-time performance for one or more specified consumers.
You can configure this action to define a Service Level Agreement (SLA), which is set of conditions that defines the level of performance that a consumer should expect from a service. You can use this action to identify whether a service's threshold rules are met or exceeded. For example, you might define an agreement with a particular consumer that sends an alert to the consumer if responses are not sent within a certain maximum response time. You can configure SLAs for each virtual service/consumer application combination.
For the counter-based metrics (Total Request Count, Success Count, Fault Count), CloudStreams sends an alert as soon as the performance condition is violated, without having to wait until the end of the metrics tracking interval. You can choose whether to send an alert only once during the interval, or every time the violation occurs during the interval. (CloudStreams will send another alert the next time a condition is violated during a subsequent interval.) For information about intervals, see the section The Intervals for Metric Publishing in the document Administering webMethods CloudStreams.
For the aggregated metrics (Average Response Time, Minimum Response Time, Maximum Response Time), CloudStreams aggregates the response times at the end of the interval, and then sends an alert if the performance condition is violated.
By default, this action does not include metrics for failed invocations. To include metrics for failed invocations, set the pg.PgMetricsFormatter.includeFaults parameter to true in IntegrationServer_directory\packages\WmCloudStreams\config\resources\wst-config.properties.
To set the Monitor Service Level Agreement (SLA) action parameters
1. In the CloudStreams Governance view, click the policy name.
2. In the policy editor on the right side of the page, double-click Monitor Service Level Agreement (SLA) in the Applied Actions list, and set the following action parameters.
Alert Interval
Number. The time period (in minutes) in which to monitor performance before sending an alert if a condition is violated.
Alert Frequency
String. Specifies how frequently to issue alerts for the counter-based metrics (Total Request Count, Success Count, Fault Count).
Every Time: Issues an alert every time one of the specified conditions is violated.
Only Once: Issues an alert only the first time one of the specified conditions is violated.
Alert Message
Optional. Specify a text message to include in the alert.
Send Data To
By default, this action logs the alerts to the CloudStreams Analytics database.
Note:
Ensure that you select the Database Publishing option in Integration Server Administrator (go to Solutions > CloudStreams > Administration > Database), as described in the section Setting the Database Options for Publishing Run-Time Metrics and Events in the document Administering webMethods CloudStreams.
Metrics Collection Level
The run-time performance metrics for a virtual service (which is invoked only in the inbound run-time scenario), are collected at the service level. That is, the metrics for all invocations of a single virtual services are aggregated together during your specified metrics publishing interval and then published.
In contrast, the metrics for a connector virtual service (which is invoked only in the outbound run-time scenario) can be collected at two different levels of metric collection:
Cloud Connector Service: Remember that a single connector virtual service can be used by multiple cloud connector services. Select this option if you want to collect the metrics for the connector virtual service broken down by each separate cloud connector service that uses it. For example, if a connector virtual services is used by three cloud connector services, then this option will collect the metrics of that service separately, broken down by each of the three cloud connector services that use it.
Connector Virtual Service (default): Select this option if you want to aggregate all the metrics for a single connector virtual service, even if it is used by multiple cloud connector services. For example, if a connector virtual service is used by three cloud connector services, then this option will collect the combined metrics for the connector virtual service by
all three of the cloud connector services that use it.
Action Configuration
Right-click the action name and click Add Action Configuration to specify a condition to monitor. To do this, select a condition Name (the metric to monitor), an Operator, and a Value for the condition. You can select Add Action Configuration multiple times to add multiple conditions. Multiple conditions are connected by the AND operator.
Name: The metric to monitor, which can be:
Availability: Indicates whether the service was available to the specified consumers in the current interval. A value of 100 indicates that the service was always available. If invocations fail due to policy violations, this parameter could still be as high as 100. That is, SOAP faults returned by the native provider or faults due to
CloudStreams policy enforcements do not impact Availability. Only errors that
CloudStreams interprets as a provider service being down will impact Availability.
Average Response: The average amount of time it took the service to complete all invocations in the current interval. Response time is measured from the moment
CloudStreams receives the request until the moment it returns the response to the caller.
Fault Count: The number of faults returned in the current interval.
Maximum Response : The maximum amount of time to respond to a request in the current interval.
Minimum Response: The minimum amount of time to respond to a request in the current interval.
Successful Request Count: The number of successful requests in the current interval.
Total Request Count: The total number of requests (successful and unsuccessful) in the current interval.
Alert for Applications
Right-click the action name and click Add Alert for Applications to specify the consumer application to which this Service Level Agreement will apply. You can select Add Alert for Applications multiple times to add multiple consumer applications.
Alert Email
Right-click the action name and click Add Alert Email if you want to send the monitoring alerts to an email address you specify in the Email ID field. You can select Add Alert Email multiple times to add multiple email addresses.
Note:
Ensure that you select the email options in Integration Server Administrator (go to Solutions > CloudStreams > Administration > Email), as described in the section Setting the Email Options for Logging Payloads and Sending Performance Monitoring Alerts in the document Administering webMethods CloudStreams.
Related Topics