Operate running Workloads¶
Preview
The Workload API is a preview feature, on by default.
Feature flag: Enable Access to Workload API Experimental Features
This section covers the day-to-day operations of a Workload, once it is running: starting and stopping, tuning runtime settings (replicas, autoscaling, resources), managing who can see and use it, and inspecting its lifecycle. For change events that swap the artifact or promote a draft, see Update existing Workloads.
| Resource | Description |
|---|---|
| Lifecycle states | States, transitions, pod-based truth, debugging, and how protons aggregate status. |
| Lifecycle actions | Lifecycle actions for Workloads—start, stop, delete, and the draft 8-hour TTL (time-to-live). |
| Runtime settings | Importance, names, runtime, autoscaling (scaling metrics), and resource configuration. |
| Sharing and access control | sharedRoles model and propagation to sub-resources. |
| REST: Operate Workloads | Start, stop, delete, settings, sharing, stats, events, and proton APIs. |
| UI: Operate deployed Workloads | Settings tab autoscaling, metadata edits, sharing, and lifecycle actions (deactivate, delete, clear stats). |
| UI: Access Workload endpoints | Endpoints tab with copy-ready invoke examples (Python, cURL) and OpenTelemetry instrumentation snippets. |
| UI: Configure Workload capacity | Capacity tab for throughput, quota reservations, entity capacity floors, and rate-limit policies. |