This version is still under development! Latest stable release is v0.3.2

Document Datastore

The Document Datastore is the software system used by STACKL to store and process documents. A document in STACKL is a file in either JSON- or YAML-format that models parts of the IT infrastructure, applications and services, and their configuration data. The datastore provides a single source of truth for application configuration data for IT environments. The datastore is interacted with through the API, which translates requests into tasks that STACKL distributes to workers that manipulate the datastore in a resilient and consistent way. All changes to documents in the datastore are made by STACKL’s document manager.

The datastore is a general module that is chosen by the user when STACKL is deployed. This makes it possible for the user to specify a variety of datastore systems, such as using a local file system, Redis, S3, and so on. The choice for a datastore depends on the desired performance such as simplicity, scalability, reliability, cost, and so on. Documents are specified in either JSON or YAML (a superset of JSON) which are lightweight and common formats to represent data, both for communication or configuration, and always stored as JSON.

Datastore design

The datastore is the single source of truth for STACKL’s documents to model, describe and automate the user’s IT environment. It performs three functions: (1) storing documents, (2) retrieving documents, and (3) running queries over its stored files. The datastore is a pluggable module, chosen during the deployment of STACKL. The pluggable module is accessed transparently through a general wrapper/interface. Each plugged technology, whether custom or third-party, provides an implementation of this interface so that documents are always interacted with in the same way, independent from the backend, and only the performance characteristics change.

The datastore is interacted with by workers during their handling of a task. The documents are versioned in the datastore through git, independently of any versioning the underlying module might offer (TODO). Any backend data storage system can be used so long as an interface is made. The interface needs to provide two things: (1) ensure that the data is converted to a JSON-format for STACKL to work with and store and (2) that STACKL can uniformly access the data, translating its data access to the real data access in the underlying system. Currently, STACKL implements a local file system and Redis as datastores. Future targets can be CouchDB, S3, MongoDB, YugabyteDB, and so on.

Last updated on February 10, 2020