When something goes wrong within your IT environment, StackState can send event notifications to alert you or your team mates. A message can be sent in the form of an email, Slack message, mobile ping or an HTTP POST request to any URL. Event notifications can contain detailed content on the trigger event and possible root cause. This guide will help you set this up.
Event notifications are triggered in response to health state change events or problem events generated in a view.
The health state of an element is derived from metrics and events data in the telemetry streams assigned to it, whereas the health state of a view is calculated based on the combined health state of elements within it. Propagated state changes can also be used to trigger event notifications, however, this can result in a lot of noise. The process to trigger a health state changed event notification is as follows:
Telemetry streams attached to an element provide metrics and events data.
A health check attached to the element reports a health state. Health checks can be synchronized from an external monitoring system or calculated in StackState based on the available telemetry streams.
When the reported health state of an element changes, a chain of state change events are generated:
HealthStateChangedEvent for the element itself.
PropagatedStateChangedEvent for all other elements that have been impacted by the element's state change.
ViewStateChangedEvent for each view containing the element. Note that this event type is only generated when a view's view state configuration criteria are met.
Event handlers associated with each view listen to the generated state change events and trigger the configured event notifications and actions.
StackState will group unhealthy components together into problems and generate problem events that track changes to them. Event handlers associated with a view can listen to the generated problem events and trigger a configured event notification or action.
Event handlers can be added to a StackState view to send event notifications and trigger actions in response to health state change events or problem events generated within the view.
You can add an event handler to a view from the StackState UI.
Select Events Settings on the left.
Click ADD NEW EVENT HANDLER.
Select the trigger event and event handler to run:
On event - the events that should trigger the event notification or automated action.
Run event handler - the event handler function that will run whenever the selected event is generated.
For health state changed events, StackState ships with event handler functions that can send an event notification via email, Slack or SMS, or POST to an HTTP webhook.
For problem events, you will need to create a custom event handler function.
Enter the required details, these will vary according to the event handler function you have selected.
Metrics and events data flow through StackState topology elements in telemetry streams. These telemetry streams are used by health checks to determine the health state of the element. For every change in health state, at least one state change event is generated. Event handlers can be added to a view to listen to state change events generated within the view and trigger an event notification or action when a configured threshold is passed.
The event types generated when an element state changes are described in the table below.
State change of an element Generated when the health state of an element changes. These events will be listed in the StackState UI Events Perspective.
Propagated state change of an element Generated whenever the health state of one of an element’s dependencies changes. These events are not visible in the StackState UI, but can be used for trigger an event notification.
State change of entire view
Generated only when the health state of a significant number of elements in a view changes. These events are not visible in the StackState UI, but can be used to trigger event notifications. Note that there may be a slight delay between the generation of a
You can add an event handler to a view to trigger event notifications or automated actions on specific state change events.
StackState groups the unhealthy components in a view into problems with a common root cause. For every change to a problem, at least one problem event is generated. Event handlers can listen to generated problem events and trigger an event notification or actions. StackState v4.3 does not include StackPacks that support notifications for problem events out of the box, however, the functionality to do this is available. You can create your own custom event handler function that will listen and react to the events generated by problems.
Changes to a problem result in the following event types being generated:
A new problem has been created.
New information is available for a problem. For example, a new root cause or contributing cause being added to the problem, or an update to a check on an existing unhealthy component.
A change in root cause has caused an existing problem to be joined with another problem.
The root cause component and all contributing cause components have reported a CLEAR (green) health state. No unhealthy components remain in the problem.
Event handlers listen to events generated within a view. When the configured event type is generated, the event handler function is run to send an event notification or trigger an action in a system outside of StackState. For example, an event handler function could send an email or make a POST to a webhook URL. A number of default event handler functions are included out of the box with StackState, or you can create your own custom event handler functions.
StackState ships with the following event handler functions that track health state change events in a view:
Event handler function
Sends a message with detailed content on the trigger event and possible root cause to the configured Slack webhook URL. See how to create a Slack webhook (slack.com).
Sends details of a health state change event using the configured SMTP server.
HTTP webhook POST
Sends an HTTP webhook POST request to the specified URL.
Sends details of a health state change event using MessageBird.
To run an event handler in response to problem events generated in a view, you will need to create a custom event handler function.