Deployment Environment Classification#

DevLake uses three regex pattern fields in per-repo scope configs to classify CI/CD pipeline runs and jobs as deployments and to assign them environment labels. These classifications feed the four DORA metrics (Deployment Frequency, Lead Time for Changes, Change Failure Rate, Mean Time to Restore). 1

The Three Pattern Fields#

All three fields are stored as plain strings in each plugin's scope config model and are compiled into a RegexEnricher at task-start time.

Key default behavior: ProductionPattern uses ReturnNameIfOmittedOrMatched — if the field is empty, every deployment is automatically labeled PRODUCTION. DeploymentPattern uses ReturnNameIfMatched, so it must match explicitly for a run to be classified as a deployment. 2

Plugin Support#

The pattern fields are present in the scope config structs for all three major CI/CD plugins:

• GitHub — GithubScopeConfig: all three fields (DeploymentPattern, ProductionPattern, EnvNamePattern), stored in _tool_github_scope_configs
• GitLab — GitlabScopeConfig: all three fields, stored in _tool_gitlab_scope_configs
• Azure DevOps — AzuredevopsScopeConfig: DeploymentPattern and ProductionPattern only (no EnvNamePattern), stored in _tool_azuredevops_go_scope_configs

How Patterns Are Applied at Extraction Time#

The RegexEnricher is built from the scope config during PrepareTaskData and stored on the task data object. Extractors then call it per-record:

• GitHub workflow runs (cicd_run_extractor.go:65-66):

  • Type ← ReturnNameIfMatched(DEPLOYMENT, run.Name)
  • Environment ← ReturnNameIfOmittedOrMatched(PRODUCTION, run.Name, run.HeadBranch) — branch name is also checked for environment matching

• GitHub jobs (cicd_job_extractor.go:87-88):

  • Type ← ReturnNameIfMatched(DEPLOYMENT, job.Name)
  • Environment ← ReturnNameIfOmittedOrMatched(PRODUCTION, job.Name) — name only

The difference: run-level environment classification also tests the HeadBranch (e.g., main, release/*), so you can scope production to a branch pattern without renaming your workflows.

RegexEnricher Utility#

enrich_with_regex.go is the shared helper used by all CICD extractors across plugins. It compiles patterns once per task run and exposes two key methods: 3

• ReturnNameIfMatched(name, targets...) — returns name if the stored regex for name matches any target; returns "" if the pattern is absent or no target matches.
• ReturnNameIfOmittedOrMatched(name, targets...) — returns name unconditionally if no pattern is registered; otherwise falls through to ReturnNameIfMatched.

Configuration Reference#

Full configuration checklist for enabling DORA metrics end-to-end is documented in the DORA Metrics Configuration knowledge page.

Key Source Files#