Transforming on-premises application for high availability

Customer Overview

One of our US-based clients has been providing technology services to large organizations since its founding in 2007. Over the years, it has grown into a major player in enterprise services. Its SaaS platform processes large volumes of data daily and serves a geographically dispersed client base. Despite its success, the platform relied on legacy components that were not high-availability aware, creating a risk of system failures. To maintain its market-leading position, the client engaged TenUp to deliver a robust, on-premises solution that would make their SaaS application highly available. This engagement focused on strengthening the client’s high availability architecture, initiating a structured application modernization effort, and optimizing their on-premises infrastructure.

Project Overview

The client needed to ensure their SaaS platform could deliver uninterrupted service to users and support real-time operations for new client provisioning. Key requirements included highly available role-based authentication, reliable backend services to prevent system-wide failures, mechanisms to avoid “split-brain” scenarios that could corrupt data, and continuous health monitoring of all node instances to preempt potential issues. This required a strong focus on failover and disaster recovery strategies as well as effective cluster management. Meeting these objectives was critical to maintaining uptime, protecting data integrity, and supporting 24/7 operations.

Challenges

Ensuring a cost-effective, seamless, and resilient high-availability solution that can integrate with the existing system, avoid downtime, prevent data issues, and include continuous health monitoring. To modernize the SaaS platform, while meeting our client’s unique requirements, the TenUp team faced the following challenges:

• Providing a solution that was easy to manage and cost-effective, ensuring long-term maintenance did not become more expensive.
• Integrating new high-availability functionality with the existing system without impacting the current codebase or disrupting business processes, while aligning with application modernization goals.
• Implementing changes within a tight timeline while avoiding performance loss or downtime during core working hours for enterprise users.
• Designing backend services to be resilient and capable of preventing “split-brain” scenarios that could lead to data corruption, supporting robust failover and disaster recovery.
• Establishing continuous health monitoring of all node instances to detect and preempt potential failures before they affect the platform, including effective cluster management.

Solution

Our application modernization efforts to deliver a highly available and resilient platform included implementing a robust architecture, combining active-passive node design, high availability architecture, high-availability clustering, and failover and disaster recovery mechanisms. The solution focused on the following key areas:

• Implemented active-passive node architecture to ensure seamless failover while keeping the system easy to manage for the client’s existing developer team.
• Used quorum witness storage to prevent “split-brain” scenarios and maintain data consistency across nodes.
• Built high-availability clusters with ClusterLab, managing active and standby node instances using Corosync for group communication and cluster management with leader-follower roles.
• Leveraged Pacemaker as a cluster resource manager to handle failover of critical resources, including floating IPs, Tomcat, and database services.
• Synchronized PostgreSQL databases across master and replica nodes using Distributed Replicated Block Device (DRBD) to ensure data consistency and minimize downtime.
• Implemented disaster recovery for all integrated components, including servers and databases, enabling rapid recovery during defined service windows.

Benefits

The implemented solution delivered significant improvements in system resilience, cost efficiency, and long-term maintainability:

• Eliminated the risk of system downtime through a high-availability, redundant architecture with no single point of failure.
• Enabled uninterrupted service delivery for enterprise users, even during node or component failures.
• Delivered a cost-effective, on-premises solution that meets availability goals without increasing operational expenses.
• Simplified system maintenance and configuration, allowing the client’s internal team to manage the platform with ease.
• Ensured scalability to support a growing user base without additional infrastructure or maintenance overhead.

Technology

• ClusterLab
• Corosync
• Pacemaker
• Distributed Replicated Block Device (DRBD)
• PostgreSQL
• Tomcat
• Quorum Witness Storage

Industry

• Enterprise Software / IT Services