Foldunk is a low-dependency set of infrastructure, examples and tests illustrating a consistent approach to Event-sourced Decision processing against ordered stream stores in F#.
The key consumer of this codebase is the Jet.com production site.
However,the codebase as a whole; particularly the, Samples are intended to achieve a critical secondary goal of sharing/discussing/prototyping usage patterns. As such, it's intended to be feasible to use it in the following contexts:
- a testbed for illustrating and prototyping decision/state processing mechanisms
- sample codebase for both internal and external facing Jet.com training events and talks
- base infrastructure for third party talks / tutorials
Please raise GitHub issues for any questions so others can benefit from the discussion.
- Foldunk.Handler: a store agnostic Decision flow runner that fulfills key request processing objectives as dictated by the requirements of Jet.com's front-end API layer
Foldunk.Serialization:Foldunk.EventStore: Production-strength EventStore Adapter instrumented to the degree necessitated by Jet's production monitoring requirements- sufficient logging and metrics instrumentation capabilties to be used in a Production context
- Does not emit any specific logs, but is sufficiently instrumented (using Serilog to allow one to adapt it to ones application (we feed log info to NLog and onwards to Splunk atm and feed the metrics embedded therein to prometheus; see relevant tests for examples)
EventSumencoding: a scheme for the serializing Events modelled as an F# Discriminated Union with the following capabilities:- Compaction: A pattern employed to optimize command processing by employing in-stream 'snapshot' events with the following properties:
- no additional roundtrips to the store at either the Load or Sync points in the flow to facilitate snapshotting
- support (via
EventSum) for the maintenance of multiple co-existing snapshot schemas in a given stream