Advanced Usage

This section covers the extension points, the integration surfaces, and the operational side of running Queue Autoscale in production. It assumes you already have the autoscaler running — start with How It Works and Configuration if you do not.

Upgrading

Upgrading to v4 - PHP 8.4, the split worker timeouts, the removed global workers block and the renamed cluster metrics.
Upgrading from v2 to v3 - Renamed APIs, the restructured config, and what the migration command actually does

Extensibility

The package has exactly two pluggable seams, plus container bindings for every internal collaborator.

Custom Strategies - Replace the demand calculation. Runs before the engine and answers "how many workers does this queue need?"
Policy Execution Internals - Adjust a finished decision. Runs after the engine, chained, with error isolation
Integrations & Developer Hooks - Facade, cluster JSON snapshot, event stream and the telemetry provider

Which seam do you need?

You want to…	Use
Change how the worker count is calculated	A strategy
Clamp, floor or veto a calculated target	A policy
React to what happened without changing it	An event listener
Export state to a dashboard or metrics pipeline	The snapshot, events, or telemetry

Production

Production Deployment Reference - Prerequisites, supervision, environment variables, what the spawned workers actually run, and the operational runbook
Security - Reporting a vulnerability, and the security-relevant behaviour of spawning and configuration

Contributing

Contributing - Development setup, coding standards and the quality gate

Prerequisites

Understand How It Works
Be familiar with the Configuration keys
Have the autoscaler running successfully in your environment