Venus (not to be confused with the planet Venus) is the secondary self-hosted server in the Dimitrium infrastructure. Its purpose is resilience rather than raw throughput: it keeps off-site copies, provides an independent observation point and can receive selected services as a cold standby node when needed.
Role
Venus receives encrypted Restic backups from Pluto and keeps a read-only operational copy of the shared Pluto–Venus journal. It has manually activated, isolated cold-standby Podman stacks available for selected sites, including Circuit and ETAI. Before an actual recovery start, the newest backup snapshot must be applied to the standby data.
The two machines monitor each other through Tailscale and SSH. This gives the infrastructure a separate network vantage point for detecting an outage and sending an alert, rather than asking the affected server to report its own failure.
Hardware and system
Venus is an Axiomtek eBOX635-881-FL x86-64 Haswell-era system with an Intel Pentium G3320TE processor, 2 GB of DDR3 SO-DIMM memory and a 120 GB SSD. It runs a Fedora-based Linux setup with SELinux enforcing and firewalld enabled.
Its storage is deliberately divided between a 30 GB root filesystem and a separate 80 GB XFS volume at /srv, with additional LVM reserve retained for future changes. The machine is modest, but its separation from Pluto is precisely what makes it useful for backups, monitoring and recovery.
Relationship with Pluto
Venus is the counterpart to Pluto. Pluto carries the primary workloads; Venus protects the ability to restore them. The pairing favors simple, understandable recovery over an unnecessarily elaborate cluster.
Monitoring and communication
Venus is not a passive disk shelf. It communicates with Pluto through the private Tailscale network and performs the same SSH reachability checks in the opposite direction. This two-way design matters: an outage can be observed by a separate machine rather than inferred only from the failing machine itself. Venus also keeps the read-only operational-journal copy received from Pluto.
Alerts and responsibility
Venus sends a scheduled heartbeat to Healthchecks as an independent confirmation that its monitoring routine remains alive. When peer checks detect three consecutive failures, the monitoring workflow sends an ntfy notification to Aleksa Dimitrijević, who maintains the two-server setup. The notification path is intentionally described only at this level: it reports the availability event without publishing endpoints, addresses, credentials or recovery controls.