Skip to content
EncryptTechBD
Home Industries Our Work Blog Pricing About Contact Discuss Your Project
বাংলা (BN)
Website & Infrastructure 5 min read

Disaster recovery checklist for ecommerce websites

A practical DR checklist covering backups, restore tests, DNS, payments, order replay, runbooks and the evidence a recovery plan needs.

E
EncryptTechBD · 0 comments
MacBook, tablet and phone arranged on a wooden desk
Photo: Wilfred Iven / StockSnap

A disaster recovery plan is not the document in the shared drive. It is the evidence that you can restore a working store, take orders and settle payments after a failure. Most ecommerce teams have a backup. Far fewer have a tested restore, a DNS plan and a way to reconcile the orders that were in flight when the site went down.

This checklist is written for the person who owns the store, not only for the hosting provider. Work through it with the developer, the finance lead and the person who answers customer messages. Each line needs an owner and a piece of evidence.

Define recovery before choosing tools

List the functions that must come back, in order: browsing and search, checkout, payment authorisation, order capture, admin access, email and SMS, and reporting. For each, agree a recovery time objective (how long it may be down) and a recovery point objective (how much data you can afford to lose). A store can often accept a slightly older product catalogue but not a missing order.

Write down the dependencies behind each function: domain registrar, DNS provider, CDN, hosting, database, object storage, payment gateway, courier API, email provider and the people who hold credentials. A recovery plan that ignores one dependency is a plan with a hidden single point of failure.

Backup checklist

  • Scope: database, application files, media, configuration, environment secrets and DNS records.
  • Frequency and retention: matched to the recovery point objective and any legal retention rule.
  • Isolation: at least one copy is immutable or offline and cannot be deleted by the production account.
  • Encryption and access: keys are held separately from the backup, with a named custodian.
  • Monitoring: a failed backup raises an alert to a person, not only a dashboard.
  • Evidence: the last successful backup time, size and restore test date are recorded.

Restore and failover checklist

  1. Provision a clean environment from infrastructure definitions, not from a half-configured server.
  2. Restore the database, application files and media in the documented order.
  3. Rebuild search indexes, queues and scheduled jobs, then confirm they are running.
  4. Install valid TLS certificates and verify redirects, canonical hostnames and payment callbacks.
  5. Load secrets from the secret store and rotate any credential exposed during the incident.
  6. Run a test order through checkout, payment, confirmation and refund.
  7. Check the admin panel, role permissions and audit log.
  8. Reconcile orders created before, during and after the outage.

Payment and order continuity

Payments are the part most plans miss. If the site is down, some customers may still complete payment at the gateway, and some orders may exist in one system but not the other. Agree how you will pull the gateway's transaction list, match it to orders, and decide whether to fulfil, refund or contact the customer.

  • Store the gateway reference on every order and keep a fallback export process.
  • Make webhook handling idempotent so a replay does not create a second order or refund.
  • Record which orders were in progress at the time of failure.
  • Define who can issue a refund and what evidence is required.
  • Keep the courier and payment provider escalation contacts outside the store.

People, access and communication

Name an incident lead and a deputy. Keep a contact tree that works when email is down, including phone numbers for the host, DNS provider, payment gateway and courier. Write customer-facing holding messages in advance for a site outage, a payment failure and a delayed delivery.

Access is part of recovery. If the only administrator is on leave, or the password manager is hosted on the same account, recovery slows down. Use a separate emergency access path, record who used it, and rotate credentials afterwards.

Runbooks and evidence

Write one runbook per scenario: hosting failure, database corruption, accidental deletion, ransomware, DNS hijack, payment provider outage and expired certificate. Each runbook should list the first five actions, the commands or console paths, the decision points and the person to call. Store it outside the infrastructure it describes.

Evidence is what turns a runbook into a plan: the last restore time, the test order ID, the reconciliation sheet, the decision log and the post-incident review. Keep the evidence with a date and an owner so an auditor, insurer or new manager can follow it.

Test schedule and common mistakes

  • Monthly: verify backup success and restore one file.
  • Quarterly: restore the full site to a staging environment and place a test order.
  • Twice a year: run a scenario with the people who would actually respond.
  • After any major change: repeat the relevant restore test.

The common mistakes are a backup that has never been restored, a database-only backup that omits media and secrets, a copy in the same account or region as production, no DNS plan, no payment reconciliation, and no named owner. Each one is cheap to find in a test and expensive to discover during an outage.

Metrics to review after a test

  • Time to restore the database and time to a working checkout.
  • Number of manual steps that were not in the runbook.
  • Orders that could not be matched to a payment after recovery.
  • Decisions that waited for someone who was unavailable.

Review the metrics with the owners and change the runbook, not only the date of the next test.

Where this fits in our work

If you need a tested recovery plan rather than a template, start with Managed Technology & Cybersecurity. The pricing page separates monitoring, backup verification and project work. For the cost view, read backup and disaster recovery testing costs.

Sources

Services that deliver this

Want this implemented properly?

Pricing →

Every engagement starts with a written scope. If a requirement is outside it, we quote it separately before starting.

Responses (0)

No comments yet. Add the first one below.

Leave a response

Your email address will not be published. Required fields are marked *.

Keep reading

More field notes