Earlier this month we finished a major infrastructure project: expanding Transloadit into a hybrid platform that combines Hetzner Cloud, AWS, and bare metal on Hetzner. The goal is simple—keep much larger fleets online at all times so we can respond to demand spikes without waiting for fresh capacity to boot.

Why Hetzner, why now?

Our workloads are elastic and we can still launch hundreds of machines in about a minute. The challenge is that some of the traffic we serve (think video on-demand encoding, or Smart CDN requests) arrives faster than even that minute can comfortably absorb, or that predictive scaling can foresee. By hosting a larger baseline fleet in Hetzner we make it economically viable to gain room to breathe while we automatically add more capacity. It buys us that minute.

How the Hybrid Cloud is structured

  • Hetnzer Bare Metal these nodes give us predictable performance on the most demanding jobs
  • Hetzner Cloud often smaller nodes and often closer to our end users
  • AWS remains in the mix for critical infra such as our primary database and to provide the instant scale-outs we mentioned earlier.

PoPs that mirror our AWS footprint

We brought new Hetzner Points of Presence online close to the AWS locations we already operate: near us-east-1, eu-west-1, and ap-southeast-1. Keeping the PoPs in close proximity preserves the latency profile our customers are used to. Components that start processing on one provider stay on that provider whenever possible, so cross-provider hops—and the latency they introduce—remain rare.

What stays on AWS

AWS continues to be our rapid response hammer. When demand shoots beyond the base fleet, we can tap into AWS capacity to absorb the burst. Legacy workloads that depend on specific AWS services stay exactly where they are, so you do not need to reconfigure anything.

What this means for you

  • Faster response under sudden traffic surges, because more capacity is already warm.
  • Lower risk of hitting scaling penalties during events with a real-time feel.
  • The same APIs, Templates, and SLAs you rely on today.
  • No data remains on Hetzner, everything is purged unless you explicitly store files (to your own storage), as was already the case with AWS.

This Hetzner expansion is the foundation for several optimizations still to come. We are already monitoring the fleet mix and will continue to tune it so that every workload lands on the most efficient infrastructure available. As always, if you have feedback or questions, we would love to hear them.