Propagation functions
StackState Self-hosted v5.1.x
Last updated
StackState v5.1
StackState Self-hosted v5.1.x
Last updated
Propagation defines how a propagated state flows from one component to the next. Propagation always flows from dependencies to dependent components and relations. Note that this is the opposite direction of the relation arrows in the graph.
A propagated state is returned as one of the following health states:
CRITICAL
FLAPPING
DEVIATING
UNKNOWN
A component's propagated state is calculated using a propagation function, which is set during synchronization.
Assigns the transparent state as the propagated state of a component.
The transparent state is pre-calculated by StackState and used for the default propagation type (auto propagation). It's passed to each propagation function script at run time in the transparentState and autoState. The transparent state is calculated as the maximum of a component's own state and the propagated state of all the components that it depends upon. For example:
CRITICAL
DEVIATING
CRITICAL
CLEAR
CRITICAL
CRITICAL
DEVIATING
CLEAR
DEVIATING
A propagation function can take multiple parameters as input and produces a new propagated state as output. To calculate a propagated state, a propagation function has access to the element itself, the element's dependencies and the transparent state that has already been calculated for the element.
The simplest possible function that can be written is given below. This function will always return a DEVIATING propagated state:
You can also use a propagation function to implement more complicated logic. For example, the script below will return a DEVIATING state in case a component isn't running:
This code works as follows:
.withId(componentId)
The componentId is passed as long and resolved
.fullComponent()
then { component -> ... }
An async lambda function where the main logic for the propagation function resides. component is the component variable, which has properties that can be accessed using .<property name>. For example, .type returns component type ID.
System parameters are predefined parameters that are passed automatically to the propagation function script at run time.
transparentState
autoState
component
The ID of the current component.
Two types of execution are available for propagation functions:
Keep performance aspects in mind when developing functions with asynchronous execution The script APIs offer super-human levels of flexibility and even allow querying standalone services. Consider extreme cases where the function is executed on all components and properly assess system impact. StackState comes with a number of StackPacks that include tuned propagating functions. Changes to those functions are possible, but may impact the stability of the system.
Running a propagation function with synchronous execution places limitations on both the capability of what it can achieve, and the number of functions that can be run in parallel. Synchronous propagation functions do, however, have access to stateChangesRepository information that isn't available if the function runs with asynchronous execution.
stateChangesRepository can be used to return:
The propagating state of an element
The number of elements with a particular propagating state
The highest state of a given set of elements
The properties and methods described below are available for use in a propagation function script.
element.name - Returns the name of the current element.
element.type - Returns type of the current element.
element.version - Returns the component version (optional).
element.runState() - Returns the run state of the current element.
element.isComponent() - Returns True if element is a component and False if element is a relation.
element.getDependencies().size() - Returns the number of dependencies.
element.getDependencies() - Returns a set of the outgoing relations (for components) or a set of components (for relations).
stateChangesRepository.getPropagatedHealthStateCount(<set_of_elements>, <health_state>) Returns the number of elements in the set that have a certain health state, for example CRITICAL.
stateChangesRepository.getHighestPropagatedHealthStateFromElements(<set_of_elements>) Returns the highest propagated health state based on the given set of elements.
stateChangesRepository.getState(element).getHealthState().intValue Returns the health state of the element.
stateChangesRepository.getState(element).getPropagatedHealthState().getIntValue() Returns the propagated health state of the element.
To manually specify a non-default propagation function, a "propagation" block should be added to the template used for topology synchronization. If the "propagation" block is omitted, the default Auto propagation will be used.
The propagation block requires the following keys:
_type - specifies that the JSON block defines a Propagation.
function the node ID of the propagation function to use. This can be obtained using a get helper.
parameter - the node ID of the propagation function’s user parameter. This can be obtained using a get helper.
Any values required for the specified argument _type.
Examples of adding a propagation function to a template can be found below:
The example template below uses a get helper to obtain the ID of the Active/active failover propagation function with the identifier urn:stackpack:common:propagation-function:active-failover. No "arguments" are specified as the propagation function doesn't include any user parameters. Arguments for system parameters are automatically provided during run time and don't need to be specified.
The template example below includes an argument that will be passed to the propagation function as a user parameter together with the standard system parameters. The "arguments" list has one argument that matches the user parameter relationType from the Stop propagation for relation type propagation function. Arguments for system parameters are automatically provided during run time and don't need to be specified.
Propagation functions can be defined and used to calculate the propagated state of a component. Some propagation functions are installed as part of a StackPack. For example, Quorum based cluster propagation, which will propagate a DEVIATING state when the cluster quorum agrees on deviating and a CRITICAL state when the cluster quorum is in danger. You can also write your own . A full list of the propagation functions available in your StackState instance can be found in the StackState UI, go to Settings > Functions > Propagation Functions
To specify a propagation function that should be used to calculate the propagated state a component, add the to the template used in topology synchronization.
You can write custom propagation functions to determine the new propagated state of an element (component or relation). The propagation function can then be used to synchronize topology.
Returns a JSON-style representation of the component. This is the same format as is obtained from the Show Json properties menu for a component, or by using a in analytics.
The precomputed .
The precomputed . This is the same as transparentState.
User parameters can optionally be defined and used in the propagation function script. The value must be provided as an argument when the function is .
(recommended)
Functions that run with asynchronous execution can make an HTTP request and use in the function body. This gives you access to parts of the topology/telemetry not available in the context of the propagation itself. You can also use the available .
See available .
- functions with either asynchronous or synchronous execution
- functions with synchronous execution only
The element properties and methods listed below can be used in propagation functions with either asynchronous or synchronous execution. Functions with synchronous execution also have access to .
You can add logging statements to a propagation function script for debug purposes, for example, with log.info("message"). Logs will appear in stackstate.log. Read how to .
The default propagation used in StackState is . If another type of propagation should be applied to a component, this must be specified in the template and applied during topology synchronization. In most cases this will be handled by the StackPack responsible for synchronization of the component.
arguments - a list of arguments to match any user parameters that the propagation function requires. Arguments for system parameters are automatically provided during run time and don't need to be specified here. For further details, see the of adding a propagation function to a template.
_type - the type of the argument. This must match the type specified for the user parameter in the .
Function with system parameters only:
Function with system and user parameters:
