Work · Case study
Rendering that scaleswithout the bill scaling.
01 / Challenge
Render jobs queued for hours at peak and left expensive machines idle at every other hour. Capacity was provisioned for the worst day of the month and paid for on all the others.
02 / Approach
We rebuilt the pipeline around a queue with autoscaling workers, defined entirely in infrastructure as code. Jobs are sharded by cost, spot capacity absorbs the bulk of the work, and a small reserved pool guarantees the deadline-critical runs.
Observability came with it. Per-job cost, queue depth and failure rate sit on one dashboard, so a regression shows up as a number before it shows up as an invoice.
03 / Outcome
Lower render spend
Peak throughput
Peak jobs clear in minutes rather than hours, and the infrastructure bill now tracks the work done instead of the capacity held.