Prefect Work Pools and Workers¶

Truthound does not replace Prefect deployment architecture. When flows move from local execution to shared infrastructure, work pools and workers remain the operational boundary for where and how a validation flow runs.

Who This Is For¶

platform engineers running shared Prefect workers
teams promoting quality flows from local development to managed execution
operators who need environment-specific routing for validation jobs

When To Use It¶

Use this page when:

a flow works locally but now needs managed worker execution
quality flows should be routed by environment, cost, or ownership
retries and concurrency must align with worker capacity

Prerequisites¶

a working Prefect deployment model
at least one saved flow or generated quality flow
a defined environment strategy such as dev/staging/prod

Minimal Quickstart¶

Create the flow locally first:

from truthound_prefect import create_quality_flow, QualityFlowConfig

quality_flow = create_quality_flow(
    "users_quality",
    cfg=QualityFlowConfig(
        rules=[{"column": "id", "check": "not_null"}],
        engine_name="truthound",
    ),
)

Then deploy it into the worker topology your team already uses. Truthound-specific configuration should stay in the block or flow config, not in ad hoc worker-only environment logic.

Production Pattern¶

Use work pools and workers as the execution boundary, not the rule-authoring boundary.

Concern	Recommended Owner
validation rules and engine selection	flow config or block
environment routing	Prefect deployment / work pool
capacity and parallelism	worker and pool settings
secrets	Prefect blocks or your existing secrets integration

Recommended checklist:

keep one deployment per environment or ownership boundary
avoid hidden worker-only overrides for rules or engine names
align retry and concurrency policy with worker capacity

Failure Modes and Troubleshooting¶

Symptom	Likely Cause	What To Do
same flow behaves differently across environments	configuration is split between the flow and worker environment	move stable config into blocks or flow config
workers saturate during large fan-out validation	concurrency policy is missing	control concurrency at the pool/deployment layer
a deployment is hard to audit	quality configuration lives outside source control	keep rules and flow definitions versioned with code