Skip to main content
Prefect collects anonymous usage telemetry to help improve the product. This page explains what data is collected, how it’s used, and how to opt out.

​
What We Collect

Prefect SDK telemetry tracks anonymous usage patterns to understand how developers use the product. We collect:
DataExamplePurpose
Eventsfirst_sdk_import, first_flow_runTrack feature adoption
Prefect version3.2.0Version distribution
Python version3.12.1Compatibility planning
PlatformDarwin, Linux, WindowsPlatform support
Architecturearm64, x86_64Build optimization
Anonymous device IDuuid4()Session correlation (not user-identifiable)

​
Events Tracked

The SDK tracks milestone events in the “Quick Start Funnel”:
  • first_sdk_import - When Prefect is first imported in an interactive terminal
  • first_flow_defined - When you define your first flow
  • first_flow_run - When you successfully run your first flow
  • first_deployment_created - When you create your first deployment
  • first_schedule_created - When you create your first scheduled deployment
Each milestone event is only sent once per installation.

​
Server Telemetry

When running a Prefect server, anonymous heartbeat data is sent every 10 minutes to help us understand server deployment patterns. This data includes:
DataExamplePurpose
PlatformDarwin, Linux, WindowsServer environment distribution
Architecturearm64, x86_64Build optimization
Python version3.12.1Compatibility planning
Prefect version3.2.0Version distribution
API version0.8.4API compatibility tracking
Anonymous session IDuuid4()Session correlation (not user-identifiable)
Session start timestampISO-8601 datetimeUptime tracking
Server telemetry is disabled automatically when PREFECT_SERVER_ANALYTICS_ENABLED=false.

​
What We Do NOT Collect

We respect your privacy. Prefect telemetry never collects:
  • User identity, credentials, or API keys
  • Code, flow definitions, or task parameters
  • Flow names, task names, or deployment names
  • File paths or directory structures
  • Environment variables or secrets
  • Error messages or stack traces
  • IP addresses or network information
  • Any personally identifiable information (PII)

​
How to Opt Out

You can disable telemetry using any of these methods:

​
Environment Variable

# Prefect-specific setting
export PREFECT_SERVER_ANALYTICS_ENABLED=false

# Industry standard DO_NOT_TRACK
export DO_NOT_TRACK=1

​
Prefect CLI

prefect config set PREFECT_SERVER_ANALYTICS_ENABLED=false

​
Automatic CI Detection

Telemetry is automatically disabled when Prefect detects a CI environment. We check for common CI environment variables including:
  • CI
  • GITHUB_ACTIONS
  • GITLAB_CI
  • JENKINS_URL
  • TRAVIS
  • CIRCLECI
  • BUILDKITE
  • TF_BUILD (Azure DevOps)
  • CODEBUILD_BUILD_ID (AWS CodeBuild)
  • BITBUCKET_COMMIT

​
First-Run Notice

The first time you import Prefect with telemetry enabled in an interactive terminal, you’ll see a brief notice: 
Prefect collects anonymous usage data to improve the product.
To opt out: PREFECT_SERVER_ANALYTICS_ENABLED=false or DO_NOT_TRACK=1
Learn more: https://docs.prefect.io/concepts/telemetry
This notice only appears once and only in interactive terminal sessions (not in scripts, CI, or non-TTY environments).

​
Local State

Prefect stores telemetry state files locally to track which events have been sent: 
~/.prefect/.sdk_telemetry/
├── device_id      # Anonymous UUID for this installation
├── milestones.json # Tracks which milestone events have been sent
└── notice_shown    # Marker that the first-run notice was displayed

​
Questions?

If you have questions about Prefect’s telemetry practices, please open an issue on GitHub or reach out to us at help@prefect.io.