Health state
StackState Self-hosted v5.0.x
This page describes StackState version 5.0.
Overview
StackState calculates and reports the health state for elements (components and relations) and views. The following health state types are reported:
Own health state - Indicates the current health state of an element based on configured health sources.
Propagated health state - Highlights potential impact resulting from other unhealthy elements in the topology.
View health state - Summarizes the health states and/or propagated health states of all elements in a view.
Changes to a health state will generate events that can be used to trigger event notifications.
Health sources
Health data in StackState can be derived from a number of health sources.
StackState health checks
StackState health checks calculate a health state based on the telemetry or log streams available in StackState and defined for a topology element. This approach opens up the possibility to use the Autonomous Anomaly Detector (AAD) for anomaly health checks.
Existing StackPacks will provide StackState health checks out of the box.
StackState monitors
The StackState monitors system compute a health state based on a configured algorithm that combines and processes the 4T data collected by StackState. Health states computed this way are bound to topology elements using health synchronization.
Existing StackPacks will provide StackState monitors out of the box.
External monitoring systems
Health data from external monitoring systems can be sent to StackState using health synchronization. In this case, the health state is calculated by an external system based on its own rules. The calculated health state is then sent to StackState as a health stream and bound to the associated topology element. This approach is useful if you have existing health calculations defined externally, or if it is not viable to send telemetry or events data to StackState and translate the health calculation rules.
Existing StackPacks will provide health synchronization out of the box.
You can set up a custom health synchronization to integrate with external monitoring systems that are not supported out of the box.
Element own health state
StackState tracks a single own health state for each topology element (components, component groups and relations) based on information available from the health sources attached to it. The own health state is calculated as the most severe state reported by all health sources configured the element. If no health sources are present, an UNKNOWN health state will be reported.
In the StackState UI, the color of an element represents its own health state. A topology element can have any of the following health states:
Green -
CLEAR- There is nothing to worry about.Orange -
DEVIATING- Something may require your attention. A badge on the component shows the number of health checks that are currently failing.Red -
CRITICAL- Attention is needed right now, because something is broken. A badge on the component shows the number of health checks that are currently failing.Gray -
UNKNOWN- No health state available.
The element will also have an outer color if it has an unhealthy propagated health state.
Propagated health state
In addition to the own health state, StackState calculates a propagated health state for each topology element (components, component groups and relations). The propagated health state is derived from the own health state of components and relations that the element depends upon.
A topology element can have any of the propagated health states listed below:
Orange -
DEVIATING- Potential impact from anotherDEVIATINGtopology element. May require your attention.Red -
CRITICAL- Potential impact from anotherCRITICALtopology element. May require your attention.UNKNOWN- No propagated health state. There is nothing to worry about.
In the StackState UI, an outer color will be shown when an element's propagated health state is calculated as unhealthy - orange for DEVIATING or red for CRITICAL.
The color of the element itself (the inner color) represents the element own health state.
The propagated health state of an element can also be found in the following places:
In the right panel Selection details tab when information about a component or relation is displayed.
In the component context menu when you hover over a component in the topology visualization.
Propagation
The propagated health state is calculated using a propagation function. Health state will propagate from one component to the next, from dependencies to dependent components. Note that this is the opposite direction to the arrows shown on relations in the topology visualization. A CLEAR (green) or UNKNOWN (gray) health state will not propagate.
Component A depends on component B. Health state will propagate from B to A.
Component B depends on component A. Health state will propagate from A to B.
Dependency in both directions. Health state will propagate from A to B and from B to A. In other words, it is a circular dependency.
No dependency. Health state does not propagate.
You can configure custom propagation functions to customize how health state affects the overall health of your systems.
View health state
When view health state is enabled for a view, it will report a health state. The view health state is calculated based on the health of components and relations within in the view.
In the StackState UI, the view health state is reported as a one of four colors:
Green -
CLEAR- There is nothing to worry about.Orange -
DEVIATING- Something may require your attention.Red -
CRITICAL- Attention is needed right now, because something is broken.Gray -
UNKNOWN- View health state reporting is disabled.
You can check the view health state in the following places in the StackState UI:
Current view - The health state of the current view is visible in the top bar of the StackState UI and also next to the view name in the right panel View summary tab. Historical health state information for a view can be seen in the timeline Health line at the bottom of the screen.
Starred views - Starred views are listed in the StackState main menu together with their health state.
All views - The health state of all views is visible on the view overview screen. Click Views from the StackState main menu.
See also
Last updated
