Building Resilience: A Cross-Region, Cross-Account Cold Backup Strategy for AWS RDS

The Client

Beckett — a B2B platform serving small- and medium-sized hobby businesses, with significant online and operational data running on AWS RDS. Their data is their business: any extended outage or data-loss event has direct revenue and trust implications. In today's digital-first world, ensuring the high availability and resilience of data is not a luxury — it's a necessity.

The Challenge

Beckett's infrastructure faced several critical pain points that hindered their disaster recovery (DR) strategy:

• Manual Snapshot Management — snapshots were taken manually, making the process unreliable and prone to human error.
• No Cross-Region or Cross-Account Replication — in the event of a regional outage, there was no mechanism to restore data from another secure location.
• Retention Challenges — without lifecycle automation, older snapshots piled up, increasing costs and complicating recovery.
• Security & Isolation Gaps — production and backup environments weren't properly segregated, risking data integrity during potential incidents.

These limitations posed serious threats to business continuity, compliance, and financial data protection.

The Solution: A Cold Backup Architecture for AWS RDS

To address these issues, SharkOps' DevOps engineer Nishant Singh designed and deployed an automated, secure, and scalable cold backup solution leveraging native AWS services. The backup pipeline operates on three axes — region, account, and time — giving the client a defensible posture for the realistic disaster scenarios that matter.

Step-by-Step Workflow

1. Automated Daily Snapshots — AWS EventBridge is configured to trigger daily RDS snapshots in the US-EAST-1 (production) region.
2. Cross-Region Replication (Same Account) — a Lambda function copies these snapshots from US-EAST-1 to US-EAST-2, enabling geographic redundancy.
3. Cross-Account Transfer (Different Account) — a second Lambda function copies snapshots to a Disaster Recovery (DR) account in US-EAST-2, converting them into manual snapshots for controlled retention.
4. Retention Management — the system automatically deletes snapshots older than 7 days in the DR account, keeping only the most recent snapshots and optimising storage costs.

Core building blocks of the pipeline:

• RDS automated snapshots in the primary region as the source of truth.
• Cross-region copy to a geographically separate region, protecting against regional disruption.
• Cross-account share to a dedicated backup AWS account, isolated from production credentials and blast radius.
• AWS Lambda orchestrators that handle copy, share, encryption, and lifecycle on schedule.
• Intelligent retention logic implemented in code — daily, weekly, monthly tiers with predictable expiry.

Architecture Flow

The detailed flow spans three AWS accounts and two regions — a Production Account in the source region, a DR Account in the destination region, and a separate isolated account that holds the final manual snapshots under stricter retention.

1. RDS automated snapshot triggers in the source region.
2. Lambda copies the snapshot to the secondary region, re-encrypting with a region-local KMS key.
3. Lambda shares the snapshot to the backup account; the backup account creates its own owned copy.
4. A second Lambda handles retention: tags lineage, computes age tier, deletes anything past policy.
5. CloudWatch alarms surface failures and SLA breaches — silence is treated as a signal, not a victory.

The Impact

The new architecture delivered tangible benefits that transformed Beckett's data resilience strategy:

• Fully Automated Backups — reduced human error and ensured consistency.
• Cross-Region, Cross-Account DR — enhanced protection against regional failures and attacks.
• Cost-Effective Storage Management — minimised unnecessary costs with efficient snapshot retention.
• Improved Security and Compliance — account isolation and controlled access.
• Confidence in Recovery — validated, restorable snapshots in a secure DR environment.

Results

• Cross-region redundancy — no single region failure can destroy the backup chain.
• Cross-account isolation — production-account compromise no longer puts backups at risk.
• Policy-driven retention — predictable storage cost, predictable recovery options.
• Operationally hands-off — the Lambda pipeline runs without daily babysitting.
• Audit-ready — every backup has a documented provenance and lifecycle.

Technology Stack

This stack uses only AWS-native primitives — managed RDS for the data layer, Lambda for automation, IAM for secure access control, and EventBridge for scheduled automation — so there are no external dependencies to maintain.

Conclusion & Key Takeaway

Backup strategy is one of those domains where what you do before the bad day determines what happens on the bad day. Backups that share fate with production — same region, same account — are barely backups. A small investment in cross-region, cross-account, policy-driven cold backups buys back a huge amount of resilience for a small operating cost.

This case study demonstrates how leveraging native AWS services can deliver a robust, scalable DR solution without adding complexity. Beckett's proactive shift from a manual backup process to an automated, secure, and regionally diverse strategy sets a strong example for other organisations aiming to strengthen their cloud resilience.