The read-only 14-day Observe trial
Connect a read-only token. Hetscale reads your history, writes a report, then records for 14 days what it would have done live.
What it does
- Reads 30 days of metric history for the servers behind your load balancer.
- Produces a day-0 report, minutes after connecting, with ranged results: capacity needed, peaks, an estimated cost range, days with CPU pinned at 100%, and a recommended starting policy. When it is ready, your organisation’s members get one email with a link to it; nothing has changed in your project.
- For 14 days, records every decision it would have taken live and every period a server was unhealthy, with the reason, prefixed "would", using the same rules as real healing and scaling. The group page sets the days so far against the day-0 report: how often it would have scaled, its busiest point against the day-0 range, and the health events.
- Built on the metrics Hetzner already records for your servers and load balancer: CPU per node, requests per second, open connections per node, target health. Signals Hetzner does not record (RAM, latency) are listed as such, so you know what the report rests on.
- After the trial: a free capacity report every 30 days.
What it does not do
It does not create, change or delete anything. It cannot: the token is read-only, and Hetzner enforces that.
- No server is created or deleted.
- No load balancer target is added or removed.
- Nothing on your servers is installed or changed.
Sample report
- CPU per node
- Requests per second at the load balancer
- Open connections per node
typical 212 · peak 604 - Load balancer target health
observed days only
- RAM
- Latency
- Errors
web-2 was unhealthy twice in the observed days; both times a server would have been started beside it. A server of yours is never replaced. Day 0 has no health history, so no health event is invented for it. CPU at 100%: 2 days in the 30-day history.
- Falkenstein
- available now
- 88% of 30 days
- Nuremberg
- out of stock now
- 61% of 30 days
- Helsinki
- available now
- 97% of 30 days
The node range above is read against what can actually be ordered. The share comes from Hetscale's own polling, not your project's API budget; until 30 days of history exist the report says "collecting".
- Scale out
- CPU above 75% for 3 minutes
- Scale in
- CPU below 35% for 10 minutes
- Cooldown
- before any scale-in
- Warm-up
- paces the next scale-out
- Minimum
- 4 servers, yours
- Maximum
- 24: the load balancer holds 25 targets, and one stays free for a replacement
- wouldcreate web-5CPU 81% for 3 min · 4 → 5
- woulddrain web-5CPU 31% for 10 min · 5 → 4
- wouldstart a server beside web-2no response for 3 min · Hetzner: running
- wouldcreate web-5CPU 79% for 3 min · 4 → 5
- woulddrain web-5CPU 33% for 10 min · 5 → 4
- wouldstart a server beside web-2no response for 3 min · Hetzner: running
Every number in a real report is a range or a count from your own data. Nothing in it is a promise. Sharing a report is opt-in and strips names and IPs.
Connect read-onlyTrial mechanics
- 14 days, one read-only token, no card.
- After the trial, polling stops. You keep a free monthly capacity report: refresh it every 30 days with a read-only token, no polling in between, no card. A refresh reads the servers the group had when its trial ended. Thirty days after the trial ends we send one email saying the next report is ready to refresh.
- Per-group Observe mode is also a permanent feature of every plan: after a failed payment or a cancelled plan, Hetscale only observes once a 14-day grace ends, and after a downgrade the Hetzner projects beyond your plan only observe. Nodes are never destroyed.
After the trial
You choose. Stay in Observe, or run one Rehearsal to see a node created and removed while you watch, then take the group Live, acting on its own or proposing each scaling action for your approval. If you do nothing, nothing happens.
Want help going Live? Write to us and we go through your policy, health check and Rehearsal with you by email.