OVH Cloud¶
This document describes how Appcircle uses OVHcloud for workloads and supporting services, and how the OVH environment is operated.
Overview¶
Our infrastructure is distributed across Google Cloud (GCP), Teknotel(Datacenter) and OVHcloud. This split is primarily driven by cost optimization and the need to keep certain auxiliary services outside of GCP.
Account and Console Access¶
OVHcloud billing is managed via a US subsidiary account. Since Appcircle is US-based, this choice helps with credit-card security and verification procedures.
Use the following URL to access the OVHcloud management panel:
- OVHcloud US Manager:
https://us.ovhcloud.com/auth/?onsuccess=https%3A//us.ovhcloud.com/manager&ovhSubsidiary=US
Note
Keep access restricted to authorized staff only. Use company-approved credential handling and MFA where available.
Networking Model¶
All OVH dedicated servers are now configured as Proxmox nodes and joined into a Proxmox cluster. Node-to-node communication runs over a Proxmox SDN / cluster network, in addition to each node's public IP. See DevOps > Proxmox for the network/SDN and firewall design.
Warning
Each node is still directly reachable via its public IP. Host hardening, firewall policies, and port exposure discipline remain critical.
Current Inventory¶
As of today, OVHcloud contains 4 dedicated servers, all configured as Proxmox nodes. VPS instances have been decommissioned.
| Hostname | Public IP | Plan | Region | Datacenter | Status | Proxmox Node |
|---|---|---|---|---|---|---|
ns31518584.ip-198-244-213.eu |
198.244.213.83 |
Advance-4 Gen 2 | eu-west-eri | eu-west-eri-a | Normal | pve5 |
ns3226830.ip-162-19-204.eu |
162.19.204.13 |
Advance-4 Gen 2 | eu-west-lim | eu-west-lim-a | Normal | pve4 |
ns3239159.ip-51-195-234.eu |
51.195.234.155 |
Advance-1 (AMD EPYC 4244P) | eu-west-eri | eu-west-eri-a | Normal | pve1 |
ns3240017.ip-51-178-216.eu |
51.178.216.77 |
Advance-1 (AMD EPYC 4244P) | eu-west-gra | eu-west-gra-a | Normal | pve3 |
All workloads previously described per-host below (Docker containers, KVM guests, agents) now run as VMs/LXC containers managed through Proxmox. For current VM/LXC inventory, network design, and security baseline, see DevOps > Proxmox and VM & LXC Security.
Agents¶
Runner agents run inside Docker containers hosted on Proxmox-managed VMs/LXC guests — not directly on the dedicated server hosts.
Agent Configuration¶
Runner configuration lives under:
/etc/agent{number}/
Example directory layout:
root@ovh-ac02:/etc/agent2# ls -l
total 16
-rw------- 1 root root 431 Nov 19 2024 ac-agent.ini # Agent configuration
-r-------- 1 root root 2320 Dec 12 2022 gcloud # GCP credentials (handle carefully)
-r-------- 1 root root 3623 Dec 12 2022 nlog.config
-r-xr-xr-- 1 root root 140 Dec 12 2022 replace.sh # Auto-update helper
Warning
The gcloud credential file is sensitive. Do not copy it out of the host. Restrict file permissions and limit access.
Agent Control Plane¶
Agents are managed via internal tooling. For operational details and implementation, see:
https://github.com/appcircleio/ac-agent-cli
Historically, the agent supported both iOS and Android. Currently, it is used only for Android.
Service Health Checks (systemd)¶
Validate which agent services and related VM scopes are running:
systemctl list-units | grep agent
Example:
agent2.service loaded active running Agent service
Note
If a legacy agent service is disabled but still present, it may be re-enabled on reboot depending on configuration. Always verify startup behavior after changes.
Per-Node VM/LXC Inventory¶
Per-host details (running VMs, LXC containers, and agent placement) are no longer tracked on this page — they are now managed through Proxmox. See DevOps > Proxmox for network/SDN design, LXC Templates for container templates, and VM & LXC Security for the security baseline applied to all guests.
VPS instances (previously used for WHM/cPanel domain hosting and smartface.io) have been decommissioned and are no longer part of the OVH inventory.
Operational Checklist¶
Use this checklist when validating OVH hosts after maintenance or reboot.
Dedicated Servers / Proxmox Nodes¶
- Validate agent services are in expected state (
systemctl list-units | grep agent) - Validate Proxmox cluster/node health (see DevOps > Proxmox)
- Validate VM/LXC guest health via Proxmox
- Confirm no legacy services were unintentionally re-enabled after reboot
Quick Commands¶
# Host overview
uname -a
lsb_release -a || cat /etc/os-release
# Service status
a) systemctl --failed
b) systemctl list-units | grep agent