29 March, 2017
Instana comes with a predefined set of health rules that are based on expert knowledge. With this release, we now allow users to define custom rules and let them trigger custom issues. Users can find the new feature in the Instana Settings under Knowledge Management → Custom Rules/Custom Issues.
- A Custom Rule defines conditions on selectable metrics. Users can choose an entity type and a metric (e.g. Host → Load), the time-window which is analyzed, an aggregation for the given time window, and a condition (> 1, <= 0.4, …). The example shows rule, that triggers when the average over 1 minute of the CPU Steal metric is larger than 20%.
- A Custom Issue binds a custom rule to one or more entities. For that, users can choose the rule they defined before (load > 1 in this example). The rule is applied to all entities that match a given filter query. In the example, the rule is only applied to hosts inside the zone us-east-1d. Whenever the rule is triggered (that means: the defined condition evaluates to true), a issue is generated. Users can customize title, text, severity and expiration of issues. If the “Triggering incident” checkbox is active, then the issue will also start an incident, which will then contain all correlated events.
- Once a custom issue is defined, it can be easily en-/disabled in the custom issues overview:
The service mapper allowed the definition of services already based on HTTP, Elasticsearch and EJB specific values. Today, we are extending the capabilities of the service mapper to include support for Docker labels. With the help of this feature, Docker labels can be used to name services. One use case for this is tracking of changes in performance and behavior of separate versions of the same component.
The following screenshot shows an example in which the Docker label com.acme.branch is used to distinguish between multiple different versions of the shop running inside a production environment.
After we introduced the container perspective with configurable grouping, we received a lot of feedback. One of the most commonly requested additions is the ability to group by user configurable container labels. This is now possible.
This dialog is available via the grouping flyout. The following example shows a container perspective grouped by the container label “com.acme.branch”.
- Go, Nodejs and Ruby are now supported under Kubernetes.
- New metric “cluster state size” on elasticsearch dashboard to make it easier to identify problems when cluster state is growing too large.
- Improved agent support for new Docker and Kubernetes versions.
- Improved compatibility with Nginx using default listen directive.
- Improve performance and stability of Java Monitoring (attach / reattach)
- MsSQL Cluster support.
- Allow search for special characters like “:” and “-”.
- Add triggering time to incidents.
- Improve handling of intermediate spans on application map.
- Reduced memory usage of PHP tracing.
- Improved MongoDB dashboard.
- Improved UI performance.
- Correctly trace Akka Http and Spray calls under high load.
- Capture Request parameters for Jersey Calls.
- Highlighting of selected tags and the selection of tags will no longer result in invalid search queries when tags require escaping.
- Agents on Big Endian PowerPC / AIX Systems running J9 Java can now connect to the backend.