Stella Ops Workflow
Stella Ops Workflow is a set of libraries that let you run an orchestration engine inside your own services, with horizontal scaling, fine-grained control and extensions.
What that means in practice
- Inside your services. The engine runs in the process of your own .NET service. There is no separate workflow server to install, operate or upgrade: your service and its database are the whole footprint.
- Horizontal scaling. Add instances of your service. They share one database: each instance claims due work with a lease, every change to a workflow instance is checked against its version, and wake-ups travel through the database or Redis.
- Fine-grained control. Workflow versions are immutable and are activated in a separate, fenced and audited step, with rollback. Long-running work survives restarts. Human tasks carry roles, assignment and deadlines, and every HTTP operation is authorized by scope and acts as the authenticated caller.
- Extensions. Storage, signalling and transports are plugins; you add your own expression functions, task permission rules and service transports through public interfaces.
Who it is for
.NET teams that build business applications with long-running processes in them: approvals, reviews, claims, onboarding, fulfilment. Processes that wait for people, call your services, branch, time out and must be auditable, inside services you already run.
How it differs
- Temporal and Camunda 8 run a separate workflow service or cluster that your code connects to; Stella Ops Workflow runs inside your service.
- Camunda 8 models processes in BPMN; Stella Ops Workflow defines them in C# or JSON, with a compile-time check.
- Elsa Workflows is also a .NET library; Stella Ops Workflow concentrates on governed versions, audited activation and task authorization.
The full comparison, dated and sourced, covers Temporal, Camunda 8, Elsa Workflows, Azure Durable Functions, Conductor, Workflow Core, Dapr Workflow, WorkflowEngine.NET, Flowable, Camunda 7 and Operaton, and describes each approach as a design choice.
About these pages
These pages are for developers who embed the engine and write their own workflows. Each page has worked examples, configuration in JSON and environment variables, typical cases, and the pitfalls to avoid. Every step you type is on the page, and every C# example compiles as shown.
Reading paths
| You want to | Read, in order |
|---|---|
| Decide whether it fits | What is different, Standard features, Comparison, Overview, Licensing |
| Run a workflow end to end in under an hour | Overview, Getting started |
| Put the engine into an existing ASP.NET Core application | Add Workflow to your project, Hosting, Persistence and signals, Configuration reference |
| Write workflows | Defining workflows, Calling services, Canonical definitions, the tutorials |
| Build a task list or a client of the HTTP API | Human tasks and the HTTP API, HTTP API reference |
| Deploy definitions and versions with an audit trail | Governed activation and publishing |
| Operate the engine | Operations, Troubleshooting, Upgrading |
| Test your workflows | Testing |
| Understand the licence | Licensing, Packages |
| Look up a type, a member or its signature | .NET API reference |
All pages
Start here
- Standard features: definitions, human tasks, calls, timers, signals, sub-workflows, forks and maps, retries, failure branches, persistence, several nodes, versions and metrics, each with a compiled example.
- What is different: the design choices that set the engine apart, each with a compiled example.
- Comparison: how Stella Ops Workflow differs from eleven other engines, dated and sourced.
- Overview: definitions, instances, tasks, the canonical form, address kinds, signals, durable execution, governed activation and the Designated Workflow count.
- Add Workflow to your project: the feed, the packages per project, one pinned version, central package management, the lock file, a signature check and the offline archive.
- Getting started: from
dotnet add package StellaOps.Workflowto a completed human task over HTTP.
Hosting and authoring
- Hosting: the ASP.NET Core composition, authentication and scopes, the actor of a task action, health and readiness, the endpoints.
- Defining workflows: the C# builder in depth.
- Canonical definitions: the JSON form, the analyzer rules, compiling a workflow to JSON.
Calling out and the API
- Calling services: the
endpoint,http,graphqlandrabbitaddress kinds, failures and timeouts. - Human tasks and the HTTP API: every route of
MapWorkflowEndpoints, with requests and responses.
Runtime behaviour
- Persistence and signals: PostgreSQL, MongoDB, Redis signalling, several nodes.
- Durable execution: exactly-once start, registered signals, external activities, maps.
- Governed activation and publishing: import, activate, roll back, audit; the publisher; remote deploy.
- Rendering: diagrams as layout JSON or SVG.
Reference and operations
- Configuration reference: every option with its default.
- Operations: retention, dead letters, health, metrics, scaling.
- Testing, Troubleshooting, Upgrading.
Licence and packages
- Licensing: how the engine counts Designated Workflows, with a plain-terms FAQ.
- Packages: which package serves which role.
Runnable material
- Tutorials: eleven short lessons, each with its complete C# code.
Other references
- HTTP API reference: the 29 routes of
MapWorkflowEndpoints, with their policies, refusals, parameters, responses and examples. - Analyzer rules: the compile-time rules WF001 to WF006, WF010, WF020 and WF021.
- .NET API reference: one page per package of the line, with its namespaces, types and members, their C# signatures and descriptions.
