Configure topology synchronizations
StackState Self-hosted v4.6.x
This page describes StackState version 4.6.
Synchronization allows you to automatically synchronize the topology of your stack to StackState based on information from such diverse systems as discovery tools, service registries, container management tools and CMDBs.
A synchronization is defined by a data source and a number of mappings from the external system topology data into StackState topology elements. The following image gives an overview of a synchronization pipeline:
Broadly speaking, the synchronization framework makes two models to turn external data into StackState internal components and relations.
- The External Topology model interprets data and turns it into a graph as the data looks outside of StackState.
- The StackState Topology model is a graph of components as viewed in StackState.
A synchronization pipeline takes data through these two models using configurable scripts. The concepts in the pipeline are explained in depth below.
The data source configures the plugin to connect to a specific instance of an external system. The Plugins section of the documentation contains more details on configurations of specific plugins. It is possible to define several synchronizations for the same data source.
To turn external data into 'External Topology', we use id extraction. The goal of this step is to take the external data and produce the following information:
- A 'type' for the component/relation for differentiation later on
- An 'externalId'. An identifier with which the element is uniquely identifiable in theexternal source
- Multiple 'identifiers'. A set of identifiers that identify the object internally in StackState.
StackState comes with some default identity extractor functions, which should be suitable for most cases.
Next up is mapping. It specifies the transformation of external system topological data into StackState domain. Mapping is defined by model element type name that is coming from external system, mapping functions and mapping functions parameters.
- Model Element Type Name - identifier of external element type, e.g. linux, hypervisor, java
- Mapper Function - Mapper function that knows how to process the data of Model Element Type
- Parameters - values for mapper function parameters, e.g. selecting template function that knows how to create specific StackState objects
- Merge Strategy - indicates the merge strategy applied in case several components form a single entity
The merge strategy is applied when multiple components get synchronized which have the same 'identifier'. At this point StackState has to decide what data to put into the resulting component, this is what the merge strategy configures. One of the following 4 options can be chosen:
- Use Mine only - Discard the other components, use just the result of the current mapping
- Use Theirs always - Discard the result of this mapping, go with the result of the other merged components
- Merge, prefer mine - Merge fields that are mergable, if not take the data from this mapping
- Merge, prefer theirs - Merge fields that are mergable, if not take the data from the other components
Mergable fields in the component are set fields (like streams and checks) and optional fields (like version, description). Mandatory fields (like name, layer) cannot be merged, for these always one has to be picked based on Mine/Theirs configuration.
Mapping functions are defined by a groovy script and the input parameters that groovy script requires. The goal of a mapper function is to process topology data from an external system and prepare parameters for a template function. A mapper function is thus likely to be plugin specific.
Template functions are defined by a JSON template and input parameters required by the template to render elements of StackState topology - mainly components or relations. When executed template functions substitutes all handlebar parameter references with values of input parameters. Template functions must define all parameters that template body refers to.
Template functions are used in cooperation with Mapping functions to create StackState topology elements. Mapper function parse topological data of external system and prepares input parameters for Template function.