Scale, Sync, Multiple Servers

[bigstusexy]

Hello,

I've looked at discussion 4875, and while it says different servers should use different config/working directories, and at this very moment I wouldn't know how to have them use the same. What I do not see directions on how to scale/sync or just run multiple servers. Sure I can prop up a new serve but I want to enter the config once and be done. I'm torn between having the stats be as a whole or per server. If this were being done inside the application, it could even be both a breakdown per server with options to show results from the whole cluster.

If this is not a feature it would be great to have it added, I didn't know to ask first and have it become or make a feature request.

[yhay81]

AdGuard Home does not currently form a cluster or share one live configuration/data directory. Each process should keep its own AdGuardHome.yaml and working directory; sharing either directory between running instances risks concurrent writes and does not create HA.

The usual setup is:

• Run two independent AGH instances, each with its own IP, config, work directory, and query-log/statistics data.
• Advertise both DNS IPs to clients (for example through DHCP/router settings).
• Treat one instance as the configuration origin and replicate supported settings to the others with AdGuardHome Sync. The official AGH README lists that project under “Projects that use AdGuard Home” and explicitly notes that such projects are third-party, not built-in.
• Back up/test the replica before the first sync, and keep instance-specific settings (listen addresses, hostnames, certificates, DHCP role, etc.) separate where appropriate.

Do not copy AdGuardHome.yaml while AGH is running: the official configuration documentation says manual edits should be made while the service is stopped because the running process writes the file and can overwrite changes.

Statistics/query logs remain per instance; there is no native combined cluster dashboard. If you need both a per-node and whole-fleet view, collect the nodes' data into an external observability/logging system and aggregate there.

So: multiple independent DNS servers + one-way configuration sync is supported operationally; shared state, clustered configuration, and aggregate statistics would be a feature request.