Customer story
How Trading Screen automated global operations and scaled with a lean team
Database refresh cycles for UAT environments were drastically reduced from roughly 72 hours down to just 4.5 hours.
About Trading Screen
Trading screen (now TS Imagine) provides SaaS-based electronic trading solutions for the global financial marketplace, supporting hedge funds, mutual funds, and brokers. Headquartered in New York City with offices in Hong Kong, Tokyo, London, Paris, Geneva, Singapore, and Frankfurt, it operates with a lean team of three Database Administrators (DBAs) across New York, Tokyo, and London.
Overcoming manual bottlenecks
TS Imagine operates a massive and complex database infrastructure featuring five distinct environments with multiple masters and regional slaves. Their architecture encompasses approximately 100 database server instances hosting roughly 200 unique non-system schemas. As they transitioned away from legacy Sybase systems and established MariaDB 10.1 as their core backbone, manual administration became unsustainable for the three-person team.
Executing standard tasks like backups, restores, and schema changes across hundreds of instances manually consumed entire workdays. Furthermore, refreshing a 150GB User Acceptance Testing (UAT) database from production using standard tools took days, severely delaying testing cycles. The lack of centralized visibility and the ever-present risk of human error—such as deploying a script to a read-only slave instead of a master—posed severe risks to production uptime.
Why MariaDB was the perfect agile foundation
To manage their sprawling infrastructure and eliminate manual bottlenecks, TS Imagine needed more than just a place to store data—they needed a highly agile, developer-friendly database foundation. They chose MariaDB because its open, flexible architecture seamlessly supported deep custom automation. Unlike rigid, locked-down legacy systems, MariaDB’s open ecosystem allowed TS Imagine to easily integrate third-party utilities (like Percona Toolkit and Graphite) and build custom deployment bots directly into their pipeline. By leveraging MariaDB as a robust, standard SQL-compliant backbone, the DBA team was able to transition from reactive maintenance to proactive engineering, ensuring high availability and procedural consistency across their massive global footprint.
Architecting for scale and resilience
TS Imagine relies on MariaDB native replication to keep data synchronized across multiple masters and regional replica servers in Tokyo, Paris, and London. Because MariaDB provided such a reliable and open foundation, the DBA team was able to confidently build custom automation on top of it:
Centralized intelligence: They deployed “DBABot,” a centralized crawler script that runs across all MariaDB instances to aggregate configurations, track schema changes, and identify highly fragmented tables for automated optimization.
Safe schema deployments: To safely handle schema changes, TS Imagine built “RosBot”, which integrates directly with Jira to automatically fetch developer scripts, establish SSH tunnels to the correct MariaDB master, and plant the scripts without human intervention.
Accelerated restores: The team revolutionized large-scale data restores by automating Percona XtraBackup. This process uses global mounts and parallel execution to refresh 150GB UAT databases across global regions, bypassing the need to generate and ship massive binlogs.
The impact
Unmatched platform flexibility:
MariaDB’s open ecosystem served as the ultimate enabler, allowing TS Imagine to successfully plug in custom bots and third-party tools that a closed, proprietary database would have rejected.
Operational efficiency
Routine DBA tasks that previously took hours or days are now handled by cron jobs and bots, allowing a three-person team to comfortably manage 100+ MariaDB instances globally.
Reduced time-to-delivery
Database refresh cycles for UAT environments were drastically reduced from roughly 72 hours down to just 4.5 hours.
Risk mitigation
Centralized script deployment and automated configuration tracking heavily insulated the production environment against typos, rogue queries, and deployment errors.
Auditable architecture
Dynamic Confluence reporting for replication topology, slave consistency, and schema change history provides immediate, effortless compliance data for financial auditors.
Future outlook
As TS Imagine continues to mature its infrastructure, the DBA team plans to replace XtraBackup with MariaDB Backup to better align with their core engine and optimize performance. Future iterations of their automation will also integrate table and index statistic copying to improve post-restore query performance and introduce automated failure alerts for the UAT refresh process.
Get started
Start building with MariaDB
One real-time database platform you can trust, for everything you run today and everything you build next.



