Virtual Wrangler

BETA

Configure the automated render wrangler to look after your jobs.

Overview

Virtual Wrangler is an automated background service that monitors your render farm and takes corrective action based on rules you define. It runs every 5 minutes and applies each enabled wrangler policy to your active jobs and tasks.

All wrangler settings apply account-wide and affect every job in the account. You can configure each wrangler independently and enable or disable them at any time from the Virtual Wrangler page.

Virtual Wrangler is a BETA feature. Settings take effect within 5 minutes of saving.

MaxTask Runtime Wrangler

Automatically kills or retries tasks that exceed a maximum runtime you define. Useful for catching stuck or runaway render tasks that would otherwise waste compute budget.

Max Runtime — the maximum number of hours a single task may run
Kill — mark the task as failed and stop it immediately
Retry — requeue the task so it starts fresh on another machine
Notify — log the event without taking action on the task

Minimum max runtime is 1 hour. Short scenes with heavy shaders can occasionally exceed 1 hour on slower machines — set the limit conservatively.

Zone Relocation Wrangler

Moves queued jobs to a different GCP zone when they have been waiting too long for resources. This handles zone-level resource exhaustion automatically.

Max Wait Time — how many minutes a job may stay queued before relocation (minimum 90)
Priority Threshold — only jobs at or above this priority level are managed

Spot To On-Demand Wrangler

When preemptible (spot) instances are unavailable and your jobs remain pending, this wrangler automatically switches them to standard on-demand instances. This trades cost savings for faster resource placement.

Wait Time — how long to wait for a spot instance before switching (minutes)
Priority Threshold — only high-priority jobs are switched to on-demand

Syncer Wrangler

Manages file synchronization failures. When a sync operation fails (e.g. a scene file upload stalls), the Syncer Wrangler applies your configured retry or fail policy.

Max Retries — maximum retry attempts before falling back to the configured action
Sync Timeout — how long to wait for a sync before marking it as failed
Retry — requeue the task for another sync attempt
Fail Task — mark the task as failed after all retries are exhausted
Alert + Retry — log a notification and also retry the task