CADET

Unity build and release automation for game developers

---

Project maintained by naishtech Hosted on GitHub Pages — Theme by mattgraham

Build Queue

Overview

Build Queue lets you enqueue multiple jobs and process them sequentially.

Each queued item stores the selected profile and operation set. This lets you stage a release run in advance and execute it without reconfiguring each job manually.

Use cases:

• Run multiple profile builds unattended.
• Schedule repeated build/publish workflows.
• Keep work moving while Unity remains responsive.

Queue Workflow

1. Configure one or more profiles.
2. Add jobs to queue.
3. Start queue processing.
4. Monitor queued, active, and completed jobs.

Recommended workflow for repeat releases:

1. Duplicate a known good profile for each target.
2. Queue all required jobs in release order.
3. Start processing and review job output as each item completes.

Queue Statuses

Common states include:

• Queued
• Running
• Completed
• Failed
• Cancelled

Pro workflows may also show publishing/notarization-specific progress stages.

Status meaning:

• Queued: waiting for execution.
• Running: currently active.
• Completed: finished successfully.
• Failed: stopped due to validation or runtime error.
• Cancelled: manually terminated by user action.

Persistence and Recovery

• Queue state is persisted across Unity domain reloads.
• Active operation context is recovered when possible after reload.

Recovery guidance:

• If Unity restarts during a queued run, reopen CADET and verify active job state before requeueing.
• If a job cannot recover automatically, mark it failed or cancelled, fix inputs, then queue again.

Queue Actions

• Start processing
• Cancel active job
• Remove jobs
• Open logs for job diagnostics

Operational note:

• Cancel only the active job when necessary, avoid clearing completed history until logs are archived.

Edition Notes

• Queue itself is available in both editions.
• Queue items that include publish/notarize operations require Pro.

Best Practices

• Keep queue jobs small and target specific, this makes failures easier to isolate.
• Use consistent naming in profiles so queue entries are easy to identify.
• Validate dependencies before long overnight queue runs.