SKAVIO / JOURNAL

RELIABILITY NOTE

Skavio Convert keeps larger jobs visible and recoverable

Production testing shows how Skavio Convert handles large files: heavy work is deliberately serialized, extra jobs enter a persistent queue, and interrupted processing can recover without duplicate delivery.

Skavio Convert keeps larger jobs visible and recoverable

Large files follow a measured path

Skavio Convert accepts files up to 2 GiB, but not every conversion has the same workload. All video conversions and PDF-to-DOCX, PNG or JPG jobs are treated as heavy jobs. That distinction matters because the service does not pretend that a demanding conversion will behave like a small image export.

One heavy slot is intentional

Testing showed that two simultaneous heavy video conversions roughly doubled the time required for each job. Skavio Convert therefore keeps one heavy-conversion slot on the current node instead of allowing parallel work to make every large job less predictable. If another heavy file is already processing, the next one waits in a persistent first-in, first-out queue.

Recovery is part of the workflow

Queue and restart-recovery testing used durable delivery checkpoints. When processing was interrupted, the affected jobs were requeued and both test jobs completed without duplicate delivery. The same testing also covered queue behavior and cleanup, giving larger-file work a defined path beyond the first processing attempt.

What this means in practice

For a large video, a PDF conversion or several files submitted together, the useful expectation is visibility rather than instant completion. Send the file, check the quoted Stars price, and use the queue position and estimated start range to understand where the job stands. Supported paid processing failures retain the automatic Stars refund path.

For larger conversions, a visible queue and one controlled heavy-work slot are more useful than an overloaded process that fails unpredictably.

Open Skavio Convert