High Query Latency: Slow dashboard performance affected user experience and the utility of the real-time data presented. For a SaaS platform, this directly impacts customer satisfaction and puts ARR at risk from potential churn.
System Instability: Frequent I/O spikes and high CPU usage created an unreliable environment, posing a risk of service degradation during peak business hours.
Scaling Limitations: With storage utilization at 70% and key components under-provisioned, the existing architecture was not equipped to support future data growth. The inability to scale confidently meant that onboarding larger clients or handling increased data volume was a business risk, placing a ceiling on potential ARR expansion.
→ Improve Performance: Reduce query latency to ensure a responsive user experience.
→ Increase Reliability: Resolve CPU and I/O bottlenecks to create a stable database environment.
→ Plan for Scalability: Reconfigure the cluster to handle future data volume and workload increases.
→ Optimize Costs: Right-size infrastructure, specifically storage I/O, to reduce operational expenses.
→ Implement Best Practices: Upgrade the cluster and apply standard practices for security, performance, and operations.
Mydbops conducted an audit of the TiDB architecture, system metrics, storage, traffic patterns, and configurations. The audit identified the root causes of latency, including TiKV congestion and suboptimal thread pool configurations.
A major version upgrade was performed in close collaboration with the Credgenics team to ensure a smooth and supportive transition. Mydbops also applied specific configuration changes to TiKV (modifying region-max-size and tuning raft/gRPC threads) to resolve core performance issues.
Mydbops provided an actionable scaling plan. This included adding a fourth TiKV node to rebalance storage, upgrading the PD node for improved cluster management, and recommending a migration to AWS Graviton instances to improve the price-performance ratio.
I/O analysis showed that provisioned IOPS for TiKV nodes were higher than peak usage. A recommendation to reduce IOPS by 33% was implemented, leading to direct cost savings.
By tuning TiDB and TiKV parameters and performing a version upgrade, average query latency was reduced from 200ms to 50ms. This made the real-time analytics dashboard more responsive.
Optimizing TiKV configurations and flow control mechanisms reduced the overall CPU load. This tuning also eliminated the I/O spikes and transport errors, resulting in a more reliable system.
Credgenics received an expert-validated plan to scale storage and compute. The plan included actionable recommendations to reduce costs through IOPS right-sizing and migration to more efficient AWS Graviton instances.
For a SaaS leader like Credgenics, ARR is the most critical business metric. The performance of their platform is foundational to customer retention and acquisition. By creating a faster, more stable real-time analytics experience, Mydbops helped Credgenics protect its existing revenue streams from churn. More importantly, the scalable and reliable platform now serves as a solid foundation upon which to build future ARR by confidently onboarding new, larger clients.
The project identified a 33% potential reduction in recurring monthly IOPS costs. The recommendation to migrate to AWS Graviton instances provides a path for further long-term savings.
The resolution of performance bottlenecks and critical errors reduced the frequency of alerts and time spent on reactive maintenance. This allowed the Credgenics engineering team to focus more on product development.
A more performant and stable TiDB cluster gives Credgenics the technical foundation to ingest more data, support more users, and expand its analytics features.
The project prepared Credgenics for future growth. They now have a scalable TiDB architecture and can consider additional ecosystem components like TiProxy for connection management as their needs evolve.
Like Credgenics, you know that latency spikes, system instability, and scaling limits are more than just technical headaches. They are direct risks to your user experience, your team's morale, and your bottom line. Constant firefighting burns out your best people and stalls innovation.
Our approach is built on partnership. We work alongside your team to provide a seamless, supportive experience, turning your database from a source of stress into a high-performance asset you can depend on.
Imagine a future where you can onboard new clients with confidence and your engineers are free to build what's next, knowing the platform beneath them is stable, scalable, and secure. That is the peace of mind we deliver.
Talk to a Database Expert Today!
.svg)
