Applicable Scenarios
- Creating or maintaining pipeline templates
- Template instantiation, batch instantiation, template updates
- Template versions, drafts, publishing, branch version management
- PAC template capabilities
- Template permission control and store template integration
Inapplicable Scenarios
- Only modifying the structure of a single pipeline model, without involving template abstraction
- Only a YAML conversion issue, without involving the template lifecycle
- Only a store plugin or ordinary pipeline execution issue
Quick Guide
- First determine which main line you are dealing with:
- Template basic structure and version system: see
reference/1-template-foundation.md - Instantiation, batch operations, permissions: see
reference/2-instance-version-permission.md - PAC templates and store integration: see
reference/3-pac-store-integration.md - Troubleshooting, events, locks, and extensions: see
reference/4-operations-extension.md
- Template basic structure and version system: see
- For template issues, first distinguish whether it is V1 or V2. For new feature decisions, complex version management, drafts, and PAC scenarios, usually prioritize V2.
- The biggest difference between templates and ordinary pipelines is not that there is "one more model", but that there are more:
- Version semantics
- Instance association relationships
- Permission and sharing boundaries
- Possible store distribution and PAC sources
- Whenever you change the template structure or template version logic, you must check at the same time:
- Template resource persistence
- Instantiation results
- Version compatibility
- The update impact on pipelines associated with the template
- Do not mix PAC templates, store templates, and in-project templates into one mental model; their sources, permissions, and version flows are different.
High-Signal Rules
- The template module is responsible for "how to reuse and distribute pipeline definitions",
pipeline-model-architectureis responsible for "what the model itself looks like" - Template versions and instance versions must be distinguished; do not equate the template version number with the pipeline runtime version
- Batch instantiation, batch updates, and store installation will all amplify template design problems; before making changes, first assess the scope of impact on instances
- PAC templates are affected by both the template system and the YAML conversion system
Key Pitfalls
- Not distinguishing V1 / V2 and directly changing template logic, resulting in changing the wrong chain
- Only changing the template itself without looking at how pipelines that have already been instantiated are affected
- Mistaking PAC template issues for ordinary template issues, or vice versa
- Using the same set of permissions and version mindset for store templates, project templates, and public templates
- For template changes, only verifying a single template without verifying batch instances, version switching, and historical compatibility
Further Reading
- Template basics:
reference/1-template-foundation.md - Instantiation and versions:
reference/2-instance-version-permission.md - PAC and store:
reference/3-pac-store-integration.md - Troubleshooting and extensions:
reference/4-operations-extension.md - When model structure is involved: also see
pipeline-model-architecture - When YAML/PAC conversion details are involved: also see
yaml-pipeline-transfer