Vahana Major Release - Release-3.3.0.0 (VRT Upgrade to Java 21)

Vahana Release Notes – Version 3.3.0.0

Release Date: 27-11-2025
Release Type: Critical Platform Optimization & Stability Upgrade


1. Executive Summary

Release v3.3.0.0 is a major platform enhancement focused on improving stability, performance, and scalability across the Vahana microservices ecosystem.

This upgrade addresses persistent production memory issues that caused unstable pod behaviour, increased latency, degraded throughput, and higher cloud infrastructure consumption.

Key Highlights

  • Java Upgrade to 21 (from Java 17):
    All seven core microservices have been migrated to Java 21, providing improved garbage collection, better memory handling, and more stable execution in containerized deployments.
  • Resolution of Resource Bottlenecks:
    Redis connection pooling has been optimized and AWS client handling has been redesigned to avoid thread starvation, memory buildup, and unnecessary heap churn.
  • Reduction in Cloud Costs:
    Optimized JVM parameters and stable memory patterns reduce aggressive auto-scaling events and unnecessary pod restarts.

2. Production Stability – Problem Statement

Production analysis revealed that several services experienced abnormal memory accumulation and were unable to release memory effectively. This triggered repeated pod restarts, declining performance, and increased infrastructure usage.

Observed Impacts

  • Stability Issues:
    Services experienced frequent outages and random OOM (Out of Memory) crashes due to excessive memory growth, high garbage collection load, and orphaned client objects remaining in heap memory.
  • Performance Degradation:
    API throughput reduced significantly under load owing to inefficient Redis connection handling. ESB threads frequently entered a waiting state, leading to blocked operations and elevated response times.
  • Higher Infrastructure Cost:
    Repeated spikes in CPU/memory usage caused unnecessary auto-scaling, resulting in an inflated number of running pods and increased cloud costs.

3. Major Platform Upgrades

3.1 Java Runtime Upgrade: Java 17 → Java 21

The following microservices have been upgraded to Java 21 to leverage modern JVM capabilities, lower GC pause times, and improve heap utilization:

  • tpg-event-manager
  • tpg-vsheduler-exe
  • tpg-vmonitoring.exe
  • tpg-vahana-dms
  • tpg-execution-sync-agent
  • tpg-esb
  • vahana-api-gateway

This upgrade ensures smoother performance under high concurrency and better containerization compatibility.


3.2 Optimized JVM Configuration

All upgraded services now use a refined JVM configuration to ensure predictable memory behaviour, improved heap stability, and reduced GC overhead.
These configurations address the memory retention patterns previously observed in production.


4. Technical Findings & Applied Fixes

4.1 ESB (Impacting API Gateway)

Analysis indicated that Redis connections were frequently stuck in a waiting state, causing ESB threads to block and resulting in memory buildup and throughput degradation.
Fix Applied: Redis Lettuce connection pooling was implemented.
Outcome: This prevents connection starvation, enhances thread throughput, and stabilizes memory utilisation under peak load.


4.2 DMS (tpg-vahana-dms)

DMS was repeatedly creating AWS client instances, resulting in excessive heap churn and increased garbage collection pressure.
Fix Applied: AWS clients are now managed using a singleton lifecycle pattern.
Outcome: This reduces object creation overhead and improves heap stability.


4.3 API Gateway

API Gateway was operating with an insufficient Redis connection pool, especially under stress caused by ESB delays.
Fix Applied: Redis pool size and Lettuce parameters were optimised.
Outcome: Improved resilience, better concurrency handling, and reduced wait times under load.


5. Expected Improvements & Business Impact

5.1 Benefits to Existing Customers

  • Improved Platform Stability:
    Memory usage is now predictable across all microservices, significantly reducing pod crashes and OOM events.
  • Lower Cloud Costs:
    With stabilized CPU and memory patterns, unnecessary auto-scaling events are expected to reduce substantially.
  • Enhanced Performance:
    Faster request processing and improved throughput due to optimized Redis handling and reduced GC interruption.

5.2 Platform Advantages for New Customers

  • Modern, Future-Ready Runtime:
    Java 21 ensures long-term support, improved security, and modern JVM performance improvements.
  • Optimized Default Deployment:
    New customers will onboard onto a pre-tuned, high-performance platform with minimal need for post-deployment adjustments.
  • Enhanced Scalability:
    Redis and Azure connection pooling enhancements enhance capacity handling during peak load and event-heavy usage scenarios.

6. Backward Compatibility & Deployment Notes

Backward Compatibility

  • No breaking changes have been introduced in this release.
  • All existing APIs, integrations, and consuming applications will continue to function without modification.

Deployment Notes

  • All seven upgraded microservices must be redeployed to apply the Java 21 runtime and optimized JVM parameters.
  • No database or schema migrations are required.
  • The release supports zero-downtime deployment through rolling updates.