Operational proof

How we operate, back up, and hand you the keys

What we usually configure by default, how we treat backups and restore, and what happens if you stop working with Blackrack. Exact terms land in your proposal and per-environment handoff.

OpenClaw Pricing Contact

Reference architecture

Three common patterns: Blackrack-managed VPS, bring-your-own VPS, and on-prem / bare metal. Channel traffic reaches OpenClaw on your server; model APIs bill separately unless you stay fully local.

Three-column diagram: managed VPS, BYO VPS, and on-prem, each showing OpenClaw and backups

What we configure by default (per agreed scope)

This is operator-grade practice, not a substitute for your own legal or compliance review for regulated workloads.

  • Baseline OS hardening: least-privilege users, SSH keys where applicable, firewall with explicit ports.
  • TLS and reverse proxy when services are exposed to the Internet; certs and paths documented in handoff.
  • Automated backups of data volumes and configuration; retention matched to risk (written down).
  • Baseline monitoring: host health, disk space, critical services; alerts to email or an agreed channel.
  • Inbound messaging allowlists where the product supports them, to reduce abuse surface.
  • Provider secrets (API keys) on the host—not pasted into chats or plain tickets.

If Blackrack winds down—or you leave

You should still have a path: data in your infrastructure account or exported, and credentials you can rotate.

  • Inventory of what runs on your VPS or server and where configuration and data volumes live.
  • Handover of access, environment files, and secret locations (or encrypted export as agreed).
  • Short runbook for start, update, and restore without depending on us.
  • If we host the machine, we schedule a migration or export window so you can copy out to another provider.

Backup and restore

Exact numbers depend on provider, data size, and maintenance windows; this is how we align expectations in plain language.

  • RPO (worst-case data you might lose): typically between daily and hourly snapshots depending on criticality—fixed in writing.
  • RTO (time to become usable again): depends on disk size and provider; Growth+ managed plans usually include restore drills.
  • Periodic restore tests on managed stacks (frequency per plan); BYO/on-prem cadence is defined in the support agreement.

Upgrade policy

We keep the OS and OpenClaw within a window we agree with you.

  • Security patches on managed hosts on a sensible cadence; maintenance windows announced with reasonable notice.
  • OpenClaw upgrades: staged in order, with a rollback plan when upstream allows.
  • BYO and on-prem can be documented handoff-only or ongoing support; pace follows the contract.
  • Larger changes (new model provider, new channels) are scoped work or a project phase.

Incident response

No fictional legal SLAs on this page—contracted SLAs appear in proposals where they apply.

  • Primary channel: the project email thread; we triage severity and next steps there.
  • Comms: we acknowledge receipt and follow up when work is in progress or we need input from you.
  • Hours: business hours (Americas) unless your plan includes on-call coverage.
  • Aftermath: short note on root cause and fixes when applicable—no vanity badges.

Sample handoff excerpt (sanitized)

What a real delivery fragment looks like (sample identifiers):

# instance: acme-openclaw (managed VPS)
# paths
app: /opt/openclaw
config: /etc/openclaw/config.yaml
env: /etc/openclaw/.env (mode 640, group openclaw)
# service
sudo systemctl status openclaw
sudo systemctl restart openclaw
# backups
last backup: see panel / log at agreed path
restore: [linked procedure in client-private doc]

Deployment providers we routinely support

Not an exclusive list—if your provider offers a standard Linux VPS with SSH, it usually fits BYO or managed.

  • Hetzner, DigitalOcean, AWS EC2, Vultr, and similar for VPS.
  • Your office or colo metal—on-prem with agreed power and networking.
  • Any VPS running a recent Debian/Ubuntu with resources sized for the workload.

Support channels

Clear expectations—without pretending we are a 24/7 NOC by default.

  • Contact: site form and email; same intake path for non-urgent support.
  • Chat: only when explicitly included in your plan (not assumed by default).
  • Timezone: Americas (Colombia / business hours) for standard responses.
  • Languages: English and Spanish depending on how we started working together.

This page describes typical practices, not a public legal guarantee or SLA. What binds is the signed contract or proposal for your environment.

Back to the OpenClaw page

Product detail, pricing context, and how we work with you.

OpenClaw