Skip to content

Improving DevOps efficiency for a global financial firm

Which microservice is silently degrading your entire pipeline?

About the company

A global financial services institution handling vast amounts of sensitive data daily through a microservices architecture, where DevOps and QA teams are jointly responsible for system stability across frequent release cycles.
Improving DevOps efficiency for a global financial firm

Industry

Financial Services, IT Services

Key challenge

Slow, late-stage performance testing with limited visibility into inter-service behaviour; issues detected too late, causing costly investigations

Stack under test

HTTPS REST APIs (client-facing services), RabbitMQ (inter-service messaging), PostgreSQL (transactional data)

QALIPSIS deployment

Nightly CI/CD execution via Gradle plugin, with step-level monitoring and email notifications

Challenges

Why can’t aggregate metrics pinpoint the degrading service?

  • Degradation was visible but the responsible service or interaction was not identifiable.
  • A single API call could trigger messaging chains and multiple database writes.
  • Investigations consumed days, delayed releases, and eroded pipeline confidence.

Solution: how QALIPSIS was used

How to test end-to-end across API, messaging, and database?

  • HTTP steps simulated client interactions against the platform’s REST APIs.
  • Messaging plugin verified that each API call produced the expected downstream messages.
  • Join operators matched each request against its messaging and database outcomes.
  • Two issues found: messages silently lost under load, and database connections exhausted at peak.
  • Root cause: consumers acked messages on receipt before async processing completed.
  • Any async failure bypassed the retry mechanism — RabbitMQ considered the message delivered.
  • Fix: ack moved to after confirmed async completion; dead-letter queue added for exhausted retries.

How to embed testing in the nightly CI/CD pipeline?

  • Scenarios versioned alongside application code and executed nightly via Gradle tasks.
  • Step-level monitoring showed per-service API, messaging, and database latency.
  • Aggregated assertions enforced service-level thresholds — any breach failed the build.
  • Email notifications ensured failures were visible to DevOps and development immediately.

Results

faster time-to-market
increased system reliability
faster test scenario development
increased cross-team visibility

Conclusion

Challenge

Slow, periodic performance testing with no visibility into which microservice degraded the pipeline, leading to costly late-stage investigations.

Solution

QALIPSIS embedded in nightly CI/CD, testing across API, messaging, and database layers with step-level monitoring and automated quality gates.

Gains

35% faster time-to-market, 50% faster scenario development, message-loss and connection-pool issues caught before production.

More use cases to explore

Accelerate your DevOps cycle with QALIPSIS.
Explore the Platform