Creating a Disaster Recovery Plan with Infrastructure as Code

Server failures, data center outages, or cyber attacks can happen at any time. By defining your disaster recovery plan as code with Infrastructure as Code (IaC), you can recreate your infrastructure within minutes during a disaster. This guide covers everything from RPO/RTO concepts to Terraform multi-region setup, automated backups to failover automation.

DR Core Concepts: RPO and RTO

Two critical metrics form the foundation of a disaster recovery plan: RPO (Recovery Point Objective) defines the acceptable data loss duration, while RTO (Recovery Time Objective) defines the acceptable downtime duration.

Multi-Region Infrastructure with Terraform

Using Terraform modules, you can define the same infrastructure as code in both primary and DR regions. This allows you to activate the DR region within minutes during a disaster.

# Primary region
module "primary" {
  source    = "./modules/infrastructure"
  region    = "eu-west-1"
  env       = "production"
  is_primary = true

  instance_count = 3
  instance_type  = "c5.xlarge"
  db_multi_az    = true
}

# DR region (Standby)
module "dr" {
  source    = "./modules/infrastructure"
  region    = "us-east-1"
  env       = "dr"
  is_primary = false

  # Fewer resources in DR (cost optimization)
  instance_count = 1
  instance_type  = "c5.large"
  db_multi_az    = false
}

Automated Backup Configuration with IaC

By defining your backup strategy as code, you can create consistent and repeatable backup processes. The following Terraform configuration automates database and file system backups.

# Automated database backup
resource "aws_db_instance" "primary" {
  identifier              = "app-db-primary"
  engine                  = "postgres"
  engine_version          = "15.4"
  instance_class          = "db.r6g.large"

  # Backup settings
  backup_retention_period = 30
  backup_window           = "03:00-04:00"
  copy_tags_to_snapshot   = true

  # Cross-region replication
  replicate_source_db     = null
}

# Cross-region replica for DR
resource "aws_db_instance" "dr_replica" {
  provider            = aws.dr
  replicate_source_db = aws_db_instance.primary.arn
  instance_class      = "db.r6g.large"
}

Failover Automation

Manual failover processes increase the risk of errors and waste time. You can configure automatic failover with DNS-based routing.

# Health check - primary region
resource "aws_route53_health_check" "primary" {
  fqdn              = "primary.example.com"
  port              = 443
  type              = "HTTPS"
  resource_path     = "/health"
  failure_threshold = 3
  request_interval  = 10
}

# DNS failover record
resource "aws_route53_record" "primary" {
  zone_id         = var.zone_id
  name            = "app.example.com"
  type            = "A"
  set_identifier  = "primary"
  health_check_id = aws_route53_health_check.primary.id

  failover_routing_policy {
    type = "PRIMARY"
  }

  alias {
    name    = module.primary.lb_dns_name
    zone_id = module.primary.lb_zone_id
  }
}

DR Testing Checklist

Without regularly testing your DR plan, you cannot be sure it will work during an actual disaster. Apply the following checklist every quarter.

For Terraform fundamentals, check our Terraform IaC guide. For backup strategies, see our Snapshot and Backup guide. For server monitoring, explore our Prometheus + Grafana guide. HashiCorp Terraform Tutorials and AWS Disaster Recovery Whitepaper are valuable additional resources.

Frequently Asked Questions

Conclusion

Make your disaster recovery plan versionable, testable, and repeatable with Infrastructure as Code. Define your infrastructure with Terraform multi-region modules, set up automated backup and failover mechanisms, and ensure your plan works with regular DR tests.