Skip to content

How Solunea SARL optimized Thaleia XL with QALIPSIS

Where exactly does your platform slow down under concurrent load?

About the company

Solunea SARL is a provider of e-learning solutions whose flagship product, Thaleia XL, enables users to design interactive e-learning modules using Excel templates, serving businesses and educational institutions at scale.
How Solunea SARL optimized Thaleia XL with QALIPSIS

Industry

EdTech, SaaS

Key challenge

Performance degradation during peak usage; monolithic architecture struggling to scale under growing concurrent user load

Stack under test

HTTPS REST APIs (platform backend), PostgreSQL (application data and usage analytics)

QALIPSIS deployment

CI/CD pipeline via Gradle plugin, with statistical assertions for automated pass/fail decisions

Challenges

Your platform is slow — but where is the bottleneck?

  • Hundreds of creators and thousands of learners overwhelmed the monolithic architecture.
  • The platform exhibited increasing latency and intermittent timeouts under load.
  • Traditional testing showed slowness but not where slowdowns occurred.

Solution: how QALIPSIS was used

How to simulate concurrent creator and learner traffic?

  • Creator persona uploaded templates, triggered module generation, and published content.
  • Learner persona authenticated, launched modules, and submitted completion events.
  • Stages execution profile replicated a training launch: steep ramp-up then sustained load.
  • Statistical assertions enforced percentile-based SLOs for automated pass/fail signals.

How to trace bottlenecks through the database?

  • Database plugin checked application records against each simulated API request.
  • Blocking operation found: module generation held the request thread during file conversion.
  • Fix: conversion moved to an async job; API returns a job reference immediately.
  • Redundant work exposed: content-launch endpoint repeated unnecessary database writes.
  • Fix: redundant write path removed; idempotency enforced at the service layer instead.

How to prevent regressions through CI/CD?

  • Scenarios versioned alongside the Thaleia XL codebase and executed via Gradle tasks.
  • Smoke tests ran in nightly builds; full load suites ran weekly.
  • Any breach of assertion thresholds failed the build automatically.

Results

increased concurrent capacity
faster response times
faster release velocity
downtime during peak usage

Conclusion

Challenge

Performance degradation during peak training launches, with a monolithic architecture struggling to scale and no visibility into where slowdowns occurred.

Solution

QALIPSIS combined multi-persona load simulation with database consistency verification and statistical assertion thresholds, embedded in CI/CD via Gradle.

Gains

55% more concurrent users, 50% lower latency, 40% faster testing cycles, and zero downtime during peak usage.

More use cases to explore

Looking to optimize your SaaS platform's performance?
Request a Demo of QALIPSIS Today!