Fleet audit

Compare key policies across every server Hostwarden knows about.

 ❯ Run a fleet audit
 ❯ Vergleiche die Policies auf allen Servern

What it compares

Hostwarden probes each host in memory/machines/ for unattended-upgrades, sshd's effective config and SSH CA trust, firewall posture, MTA, the network stack and resolver, time sync, auto-reboot behaviour and needrestart's restart mode, mesh VPNs and their SSH servers, accounts and sudo rules and, on Ubuntu, Pro/ESM coverage.

• The result is a side-by-side table that highlights where servers disagree, with sections for drift, warnings, what you decided, and the hosts whose memory is stale or never onboarded.
• Alpine, FreeBSD and macOS have probes of their own; a setting a family does not have reads n/a, not drift.
• An appliance is compared only with its own kind.
• Windows hosts are skipped.

How it runs

In Claude Code each host gets its own subagent and returns one comparison row, which keeps the raw output out of the conversation; elsewhere it probes one host after another. Each guest stands right after the hypervisor its Runs on: line names, whether it has SSH of its own or is reached through that hypervisor; a VM in the cloud or on a host Hostwarden does not manage stands on its own. In a container, the time sync, the uptime and a kernel waiting for a reboot belong to its host and read n/a; the host's own column shows them where the host is audited.

It makes no configuration changes on any host — it only writes one audit-trail line to each journal. Use it after fixing a config bug on one server to find which others carry the same bug, or as a periodic consistency check.