WildFly & JBoss EAP: application-server expertise

By Cédric Barme · Founder of NSY

WildFly and JBoss EAP carry a considerable share of the critical systems in French banking and insurance. Cédric Barme has worked on these application servers since 2012: fine-grained tuning, high availability, major version upgrades and cloud moves — in environments where downtime is measured in regulatory exposure, not just minutes.

Typical engagements

• Major version upgrades — JBoss EAP 6/7 → 8, or EAP → WildFly moves: compatibility assessment (deprecated APIs, Jakarta namespace), configuration migration (standalone/domain, profiles, subsystems), non-regression testing and controlled cutover.
• High availability — clustering, session replication, HA JMS queues (see the messaging expertise), load balancing and failure tolerance tested under real conditions.
• Tuning and stability — JVM sizing (GC, heap, threads), connection pools, cascading timeouts; diagnosing the leaks and slow degradations that end up as production incidents.
• Containerisation — WildFly on OpenShift/Kubernetes while preserving Java EE constraints: sessions, XA distributed transactions, startup ordering (OpenShift expertise).

A server that "runs" versus a server that is mastered

Many JBoss installations run on inertia: the original configuration was never revisited, and default settings absorb the load until the day they don't. Mastery means knowing why each subsystem is configured the way it is, having metrics that warn before the incident, and upgrading without holding your breath. That is the state NSY leaves platforms in at the end of an engagement — documentation and skills handover included.

The regulated context is not optional

In banking and insurance, touching the application platform means working under constraints: auditable changes, segregated environments, negotiated maintenance windows, ACPR/AMF requirements and now DORA (see the DORA support). NSY has worked within these frameworks since 2018 — the procedures are not an obstacle discovered mid-flight, they are built into the plan from scoping.

How to start

The most effective format: a short audit of your installation (configuration, versions, metrics, recent incidents) leading to a prioritised action plan. The FAQ gives the ballparks; for a specific context, the contact form gets a reply within 48 business hours.