VegaFlow concepts

QuickFlows, pipeline projects, versions, environments, runs and operational evidence.

VegaFlow has two execution models. They share connections, managed compute, permissions and run evidence, but they solve different problems.

Use	QuickFlow	Pipeline project
Primary job	Direct source-to-destination synchronization	Versioned data, application or AI workflow
Shape	One source, one destination, many selected streams	One or more pipelines containing a task graph
Authoring	Connection discovery and mapping form/API	Source repository or development revision plus pipeline declarations
Change history	Immutable definition versions	Immutable source revisions, pipeline versions and deployments
Recovery	Retry failed streams	Retry or restart a pipeline execution according to dependency state
Triggering	Run now or cron schedule	Manual run, API schedule, event trigger or repository reconciliation

Resource hierarchy

Workspace
├── Connections
├── Compute environments
├── QuickFlows
│   ├── Definition versions
│   ├── Schedules
│   └── Runs → stream runs → attempts
└── Pipeline projects
    ├── Repository + definition roots
    ├── Source revisions
    ├── Environments → deployments
    └── Pipelines → versions → executions → runs/logs/timeline

Definitions and executions

A definition says what should run. A version freezes that meaning. An execution records one attempt to run a selected immutable version in a selected environment.

Editing a QuickFlow or project never rewrites historical runs. This is the central rule behind repeatable retries, auditability and VegaGraph lineage.

A connection stores non-secret configuration and references managed credentials. Its connector defines supported roles, configuration schema, discovery behavior and sync modes. Flow definitions reference the connection ID; they do not copy passwords or tokens.

Compute environments

A compute environment supplies execution capacity and network placement. It can stop, start or resize independently of definitions. Runs resolve the current authorized environment configuration before work starts.

Project source and deployment

A pipeline project can follow a Git repository or create a development revision. Source reconciliation resolves a commit/configuration generation, builds an immutable source revision and discovers pipeline declarations. Publishing creates immutable pipeline versions. A deployment selects what a project environment should run.

Schedules and triggers

A schedule or trigger points to an operational pipeline identity and published version through the environment deployment. It creates an ordinary execution; it does not bypass permissions, version resolution or environment readiness.

Operational evidence

VegaFlow records parent execution state, child run state, attempts, logs, timeline entries, trigger identity and immutable version references. VegaGraph can connect that evidence to datasets, code, owners, policies, infrastructure and business outcomes.

Choosing the right surface

Use a QuickFlow when the desired result can be stated as “copy these streams from this connection to that connection.” Use a pipeline project when you need transformations, branching dependencies, code review, development sessions, environment promotion, schedules/events tied to a release, or a durable multi-step workflow.

On this page