Feature Request
Describe the feature you'd like
- A YAML-based pipeline system that chains
download, deploy, and delete steps into a single sequential workflow, run via databusclient workflow run pipeline.yml. Each step's output can be referenced by later steps using ${steps.step_name.output_files} syntax, resolved at runtime.
- Each step declares its own on_error behavior: fail, continue, or retry (with configurable max_attempts/delay_seconds).
- Environment variables (${VAR_NAME}) are substituted at parse time, with a clear early error if a referenced variable is missing.
Why is this feature important?
Today, running a multi-step Databus operation (e.g. download → deploy → delete old version) requires manually chaining CLI commands, with no structured record of what was run and no way to reuse a pipeline definition. This closes that gap, building on the existing stateless download()/deploy()/delete() functions without modifying their internals.
Additional context
Depends on the manifest system (#64) and replay (#65) already being in place, since a workflow run reuses the same underlying API functions.
Feature Request
Describe the feature you'd like
download,deploy, anddeletesteps into a single sequential workflow, run viadatabusclient workflow run pipeline.yml. Each step's output can be referenced by later steps using ${steps.step_name.output_files} syntax, resolved at runtime.Why is this feature important?
Today, running a multi-step Databus operation (e.g. download → deploy → delete old version) requires manually chaining CLI commands, with no structured record of what was run and no way to reuse a pipeline definition. This closes that gap, building on the existing stateless download()/deploy()/delete() functions without modifying their internals.
Additional context
Depends on the manifest system (#64) and replay (#65) already being in place, since a workflow run reuses the same underlying API functions.