Running autonomous AI coding agents (OpenCode, Claude Code, VS Code Server) with parallel test suites, TypeScript language servers, and headless browsers requires high single-core speed (2295 GB6) and 16 GB RAM.
opencode-cli + code-server + docker-in-docker + playwright-runnerPartner notice: we may receive a commission when you order through our links. The price you pay does not change.
8 Dedicated Cores (AMD EPYC 9645) · 16 GB RAM · 256 GB NVMe · 2.5 Gbit/s
GB6: 2,295/13,420 · FIO 4K: 192,000 IOPS · Steal: 0.0 %
6875nc1792797911936nc179270376304 Dedicated Cores (AMD EPYC 9645) · 8 GB RAM · 128 GB NVMe · 2.5 Gbit/s
GB6: 2,290/7,559 · FIO 4K: 187,300 IOPS · Steal: 0.0 %
6874nc1792703772036nc17927037630Recommended kernel parameters, memory management, and security hardening for maximum reliability.
Add a 4–8 GB swapfile and set "sysctl vm.swappiness=10" to prevent the Linux OOM Killer from terminating container processes during memory allocation spikes.
If your app sends notification emails: delete the "netcup Mail block" rule in SCP Firewall and configure a clean reverse DNS (rDNS/PTR) record matching your domain.
Mount application state into named Docker volumes. Automate daily offsite backups using BorgBackup or Restic — remember SCP snapshots expire automatically after 30 days.
Keep UFW active allowing only ports 80 and 443 publicly. Shift SSH from port 22 to a custom high port to eliminate automated credential stuffing scan noise.