Skip to main content
Version: 0.9.12

Server groups

platform v0.9.11verified 2026-05-14

Path: /admin-settingsServer Groups tab

Server groups define pools of voice processing servers. They control how calls are distributed across servers (load balancing) and how the platform automatically adds or removes instances based on demand (auto-scaling).

Server group

FieldRequiredDescription
Group NameYesUnique (e.g. premium, standard, us-east).
DescriptionNoNotes.
Servers CountAutoCurrent instances.
EnabledYesMaster toggle.

Load balancing

FieldRequiredDefaultDescription
LB MethodYesround_robin / least_connections / weighted_round_robin.
Max Concurrent CallsYes100Per-server cap. When reached, new calls route elsewhere.
PriorityYes1 – 10 (higher = preferred). Used when multiple groups are available.

Auto-scaling

FieldRequiredDefaultDescription
Scaling EnabledYesfalseMaster toggle.
Min InstancesYes1Never scales below.
Max InstancesYes10Never scales above.
Scale Up ThresholdYes0.7Utilisation (0 – 1) triggering scale-up.
Scale Down ThresholdYes0.3Utilisation triggering scale-down.
Cooldown SecondsYes300Min wait between scaling actions.
Scaling ProviderYeshetzner, aws, custom.
Scaling ConfigNoProvider-specific JSON (machine type, region, SSH keys, …).

How scaling works

  1. The Scaler monitors average utilisation per group.
  2. When utilisation > Scale Up Threshold, a new instance is provisioned.
  3. When utilisation < Scale Down Threshold, an instance is removed.
  4. The Cooldown prevents rapid up / down churn.
  5. Instance count stays between Min and Max.

Chain to teams

Server Group → Base Number → Team → Endpoint. Assign groups to base numbers (Base numbers) and to teams (Users and teams).

Workflows

Create a server group

Click Create → name + description → configure LB (method, max concurrent, priority) → configure auto-scaling if needed (enable, thresholds, min/max, cooldown, provider, config) → enable → save → assign in Base numbers and Teams.

Adjust scaling thresholds

Edit the group → adjust Scale Up / Down Threshold and Cooldown → save.

See also