We asked Drew Piland, product manager at CloudBees, how the company’s solution helps organizations facilitate tasks that need to be done inside the CI/CD pipeline. This is what he had to say:
There are a few different task types: integrations with other tools, scripts, deployments, and manual tasks. Integration steps will interact with the third-party services using API calls. Scripts, on the other hand, will run on an agent. You can target a specific agent or a pool of agents with the required dependencies installed.
Deployment tasks will often run on the target deployment environment, optionally using a utility agent to offload any heavy lifting. And manual tasks will notify the assigned person of their task, optionally including a set of instructions. When a task is run, the data from the event is stored and can be referenced from any task further in the pipeline. This allows you to build automated workflows based on data between different services.
You group the tasks into stages with entry and exit gates determining when the pipeline can run the tasks within the stage. The gate rules can evaluate data from any task, such as the results from a security scan. Or it can be a manual approval where a particular person or team is tasked with approving it.
Once you build your ideal pipeline, it can easily be shared and reused by exporting it into DSL code and adding it to the self-service catalog. With a service catalog full of content, you can provide the organization with a set of compliant workflows where anyone with permissions can generate their instance of a workflow.