JBoss can be improved significantly, especially regarding deployment overlays that need updates to apply quick fixes or environment-specific changes without redirecting the archive. Enhancements in CLI scripting for automatic deployments or rollbacks and having an automated way of updating in the future for pack changes and version transitions are critical. More details about the update progress of package installations, artifacts downloading, as well as automation for modules and configurations applied are needed. Additionally, features for drifting and reverting would be worthwhile. We leverage domain mode with server groups to enforce synchronized rollout strategies across clusters without downtime or config drift.
Subsystem isolation doesn’t just reduce memory usage it allows us to apply GC tuning and diagnostic tracing at the service level, not the container level.
Our CLI scripting has been backed by pre-validated batch execution pipelines, eliminating human error during hotfix rollouts and version transitions.
Using overlays in conjunction with marker files, we've created an audit-friendly patching workflow that doesn't require full archive redeployments.
Changes in the future need to align with today's directions regarding the most evolutionary topics of Jakarta EE progression. As Jakarta EE progresses, newer specifications such as Jakarta Data and Jakarta NoSQL or AI-assisted diagnostics are necessary.
We’ve integrated JBoss metrics output with Prometheus exporters, enabling real-time subsystem-level observability and predictive scaling alerts.
By aligning with Jakarta EE's modular progression, we've positioned our stack to adopt emerging specs like Jakarta NoSQL without disruptive upgrades.
JBoss’s flexible threading model allowed us to apply workload-specific executor policies, preventing starvation in high-concurrency deployments.
More straightforward updates and rollbacks need to be done with the CLI, alongside improved observability, such as native support for OpenTelemetry or enhanced DevOps tools with command-line interfaces and automation features. Support for YAML-based configuration is crucial, especially in a GitOps deployment style, along with cloud-native enhancements such as integration with Kubernetes, OpenShift, or newer technologies.
I would like to see JBoss reassess its executor configuration controls and consider offering default workload profiles such as I/O-bound, CPU-heavy, or async-first—to optimize threading strategies out of the box.
On the cloud-native side, it’s important to validate container readiness and expand operator-driven automation for Kubernetes, especially focusing on CRD evolution and stateless rollout support. I would also recommend improving workflow transparency by providing clearer feedback during pack updates, including artifact download status, config sync logs, and rollback outcome visibility.