CI/CD Providers
Choose your platform for credentials, discovery, runnable instrumentation, and verification.
Capability Matrix
| Provider | Sensor detection | Team Server discovery | Result |
|---|---|---|---|
| Bamboo | Yes | Yes | Managed protection after plan discovery |
| GitHub | Yes | Yes | Managed protection after repository discovery |
| GitLab | Yes | Yes | Managed protection after project discovery |
| Jenkins | Yes | Yes | Managed protection after job discovery |
| Koji | No | Yes | Inventory only; builds remain unobserved |
| TeamCity | Yes | Yes | Managed protection after project discovery |
Discovery creates inventory; runtime detection reconstructs the build’s identity. Managed policies require both and an exact existing pipeline match.
Detection-only Providers
| Provider | Detected identity |
|---|---|
| AWS CodeBuild | Project and build |
| Azure Pipelines | Project, definition, job attempt, and build; each attempt has its own job ID |
| CircleCI | Persistent project/repository, job, and build; project ID identifies the pipeline across runs |
Detection supplies metadata and cancellation handling, not pipeline creation. Without an exact existing pipeline, these providers use local policies; Team Server does not save their remote build records or violations.
Manage Discovery Integrations
Creation and update validate credentials before saving and queue combined pipeline/build discovery. Scheduled jobs continue discovery. Start Sync requests a combined run; Stop Sync is available while Syncing.
Statuses are Active, Syncing, Stopping, and Error. Error corresponds to API inactive: a required check or discovery stage failed. Deleting an integration removes credentials, configuration, and run history, retaining discovered pipelines/builds.
The table shows the last successful run’s time, duration, and scope, plus the next provider-wide schedule. Integrations for one provider share that next schedule. Failed/partial attempts remain in history without replacing last-successful values.
Review Run History
Operators and Administrators can View Run History. Filter/search by outcome, scope, or trigger, order by start time, paginate, and export CSV. Triggers include manual, scheduled, create, and update.
| Outcome | Meaning |
|---|---|
| Running | In progress |
| Success | All requested stages succeeded |
| Partial | Completed with recoverable failures |
| Failed | Required stage failed |
| Cancelled | Operator stopped it |
| Interrupted | Server stopped before finalization |
Diagnostics shows stage, severity, timestamp, and message; the drawer reports omitted records when its limit is reached. Retention defaults to 90 days under Administration > Log Settings, enforced by log_prune confirm:true.
Provider Mapping
One pipeline is one policy-assignment boundary:
| Provider | Pipeline | Build | Policy covers |
|---|---|---|---|
| Bamboo | Plan | Job result | Every job in the plan |
| GitHub | Repository | Workflow run | Every workflow/job in the repository |
| GitLab | Project | CI job | Every project job |
| Jenkins | Executable job | Build | That job |
| Koji | Package | Unique binary RPM architecture, or one build without binary architectures | Inventory only |
| TeamCity | Project | Build | Every project build configuration |
Organizations, groups, Bamboo projects, Jenkins folders, and multibranch parents scope discovery rather than becoming pipelines. Executable Jenkins children become separate pipelines. TeamCity subprojects containing build configurations are separate pipelines.