A Recovery Plan That Brings Data, Applications and Traffic Back in Order
Ficode configured an AWS active-passive recovery workflow with a European primary region and a warm standby in APAC. Health is checked at one-minute intervals; eight consecutive failures trigger further confirmation and readiness checks. The workflow then coordinates data promotion, application validation and traffic switching.
The project in context
The client operates a specialist veterinary healthcare platform with continuous clinical updates, uploaded images, notifications and shared records. Ficode configured an AWS active-passive recovery workflow with a European primary region and a warm standby in APAC.
Continuous updates. Interdependent services. A recovery sequence too important to improvise.
A regional outage could interrupt a write-heavy clinical platform while records, uploads and notifications were still changing. Restoring service required more than starting new servers: data state, application dependencies and external traffic had to be addressed in the right order.
The challenge Before
Clinical activity
Continuously updated recordsRecovery dependencies
Data and application must be readyTraffic decision
Regional switch needs safe checksWith Ficode Connected Approach
Confirmed outage
Monitor and reconfirm failureValidated standby
Promote data; check applicationControlled switch
Global Accelerator dials and statusCoordinate recovery through checks, promotion and traffic control
Detect
CloudWatch-based monitoring checks the primary region every minute. After eight consecutive failed checks, the workflow starts and reconfirms the outage.
Prepare
The workflow checks APAC standby readiness before promoting DocumentDB, RDS for PostgreSQL and Redis through ElastiCache.
Validate
Following data-service promotion, the workflow validates application endpoints and load-balancer target groups.
Switch
AWS Global Accelerator traffic dials shift traffic from the European primary towards APAC. SNS handles notifications; DynamoDB records workflow state.
From primary-region failure to a controlled standby traffic switch
Detect and confirm
1-minute checks; 8 failures; reconfirmPrepare and validate
Standby readiness; data; applicationSwitch and record
Traffic dials; SNS; DynamoDBCONNECTED OUTCOME
An ordered, gated recovery sequence brings the APAC standby into service after data and application checks.
Recovery coordination with clear operational boundaries
A defined outage trigger
One-minute health monitoring and eight consecutive failed checks provide an explicit trigger.
Data and application checks before routing
The workflow checks standby readiness, promotes data services, then validates application endpoints before moving traffic.
Controlled regional traffic switching
Global Accelerator traffic dials provide the configured route from the European primary region to the APAC standby.
Traceable recovery coordination
Step Functions and Lambda coordinate recovery actions, with SNS notifications and DynamoDB state tracking.
This engagement established a configured disaster-recovery workflow. No live end-to-end recovery time or data-loss result has been measured. RTO up to 60 minutes and RPO up to 15 minutes are design objectives.
Confirmed technology and implementation scope
CloudWatch, EventBridge, Lambda, Step Functions; DocumentDB, RDS, ElastiCache, Global Accelerator; SNS, DynamoDB, S3.
