Hetzner load balancer, two or more servers, registry image

This is the case Hetscale was built for, and the one with no prerequisite work: the load balancer’s health check already decides readiness, and the image already makes every server the same. Fit 10; the setup is the 15 minutes Hetscale needs in every case.

Prerequisite work
none
Time to set up
15–20 min
Adapter
Snapshot or webhook
Fit
10/10

Prerequisites, time and fit are the ADR-0028 row for this case. Hetscale's own setup is about 15 minutes in every case; the rest is work on your side.

Steps

  1. 01
    Connect read-only

    Create a read-only Hetzner token for the project and start the 14-day Observe trial. Hetscale reads 30 days of metrics for the servers behind your load balancer and writes the day-0 report.

  2. 02
    Pick the template

    Either point Hetscale at one of your servers as the seed and let it snapshot it, or use the webhook adapter so your pipeline deploys the image to each node Hetscale creates.

  3. 03
    Run one Rehearsal

    One node is created from the template, deployed, health-checked, drained and deleted while you watch. Success unlocks Live.

  4. 04
    Go Live

    Set the range. With two servers already serving, minimum 2 keeps zero-downtime healing.

FAQ for this case

Does Hetscale touch my existing servers?

No. They are listed, never modified or deleted; only servers Hetscale created carry its labels and can be removed by it.

Which server type do new nodes use?

The group’s template: server type, locations, network, firewall, SSH keys. Architecture is locked to the seed; a fallback type or location you allow is used when the preferred one is out of stock.

See what Hetscale would have done with your real data — connect read-only, get your report in minutes.

Connect read-only