Explore Prefect concepts¶
Concept | Description |
---|---|
Flows | A Prefect workflow, defined as a Python function. |
Tasks | Discrete units of work in a Prefect workflow. |
Deployments | A server-side concept that encapsulates flow metadata, allowing it to be scheduled and triggered via API. |
Work Pools & Workers | Use Prefect to dynamically provision and configure infrastructure in your execution environment. |
Schedules | Tell the Prefect API how to create new flow runs for you automatically on a specified cadence. |
Results | The data returned by a flow or a task. |
Artifacts | Formatted outputs rendered in the Prefect UI, such as markdown, tables, or links. |
States | Rich objects that capture the status of a particular task run or flow run. |
Blocks | Prefect primitives that enable the storage of configuration and provide a UI interface. |
Task Runners | Configure how tasks are run - concurrently, in parallel, or in a distributed environment. |
Automations | Configure actions that Prefect executes automatically based on trigger conditions. |
Block and Agent-Based Deployments | Description |
---|---|
Block-based Deployments | Create deployments that rely on blocks. |
Infrastructure | Blocks that specify infrastructure for flow runs created by a deployment. |
Storage | Lets you configure how flow code for deployments is persisted and retrieved. |
Agents | Like untyped workers. |
Many features specific to Prefect Cloud are in their own navigation subheading.