Backup and Disaster Recovery

Many businesses discover their backups are incomplete on the day they need them. The plugin stopped running months ago, the database was excluded, or the only copy sat on the same server that failed. CISA's ransomware guidance recommends keeping offline, encrypted backups of critical data and regularly testing their availability and integrity. We build backups around that idea.

The 3-2-1 approach

• Three copies of important data: the live copy and two backups
• Two different storage systems, so one failure cannot take out both
• One copy off-site, in a separate location and account from production

Off-site copies are encrypted before they leave the server, and the backup storage uses separate credentials so an attacker who reaches the website cannot delete its backups. Where possible, backups are kept in an immutable or versioned form for a retention period.

• Website files, themes, plugins and uploaded media
• Databases, including continuous backups for busy applications
• Store orders, customer accounts and product data
• Self-hosted business software such as CRM, help desk and e-signature data
• Configuration, DNS records and deployment settings needed to rebuild

Backup frequency and retention are set by how much data you can afford to lose and how far back you may need to go. A store taking orders all day needs more frequent database backups than a brochure site that changes once a month.

On a schedule, we restore backups into a separate environment and check that the site loads, logins work and recent data is present. Results are recorded so you have evidence, not assumptions. We also time each restore, so the recovery plan is based on how long a restore really takes rather than a guess. Every backup job also reports in when it finishes, and an alert fires if a job fails or does not run.

A recovery plan answers practical questions before an emergency: who decides to restore, which systems come back first, where the restored site will run, how DNS gets switched and who tells customers. We write the plan with you, agree on recovery time and data loss targets for each system, and keep a copy outside the systems it describes. After ransomware, a hacked site or a failed server, the plan turns a crisis into a checklist. We review it at least once a year and whenever a major system changes, so it never describes servers that no longer exist.