Runs & Jobs
Ce contenu n’est pas encore disponible dans votre langue.
A run represents a single execution of a pipeline. Each run contains one or more jobs that agents execute. The Orchestrator provides real-time monitoring, log streaming, and artifact collection for every run.
🔄 Pipeline Execution Flow
Section titled “🔄 Pipeline Execution Flow”- A run is triggered (manually, via API, or by webhook)
- The orchestrator queues the run and creates jobs
- An available agent matching the job’s required labels claims the job
- The agent downloads the pipeline, executes graph nodes, and streams logs
- On completion, artifacts are uploaded and the run is marked as succeeded or failed
▶️ Triggering a Run
Section titled “▶️ Triggering a Run”From your project page, click the Run button and select a pipeline and branch. The run enters the queue and is picked up by the next available agent.
For API and webhook triggers, see Triggers and REST API.
📊 Monitoring Runs
Section titled “📊 Monitoring Runs”Run Details
Section titled “Run Details”Click on any run to see its detail page:
- Status: Queued, running, succeeded, failed, or cancelled
- Jobs: Individual job statuses within the run
- Duration: Total execution time
- Attempt number: Runs can have multiple attempts (re-runs)
Live Logs
Section titled “Live Logs”The log viewer streams output in real-time while the agent executes. It supports ANSI colors for formatted terminal output. Logs are preserved after the run completes.
Graph Viewer
Section titled “Graph Viewer”For graph-based pipelines (.act files), the run detail page embeds an interactive graph viewer showing which nodes are currently executing, making it easy to follow the pipeline’s progress visually.
📋 Queue Management
Section titled “📋 Queue Management”The Queue page shows all pending and running jobs:
- Drag-to-reorder: Prioritize pending jobs by dragging them in the queue
- Bulk cancel: Select and cancel multiple queued runs
- Real-time updates: The queue updates live as jobs are claimed and completed
🔁 Re-running
Section titled “🔁 Re-running”On the run detail page, click Re-run to create a new attempt. All previous attempts are preserved and accessible from the run page.
📦 Artifacts
Section titled “📦 Artifacts”Agents can upload artifacts during job execution. Once a run completes, artifacts are listed on the run detail page with:
- Filename and size
- SHA-256 checksum
- Download link
The Artifacts page provides a global view with search and filtering across all runs and projects.
For programmatic access to runs and artifacts, see REST API.