Appearance
Deployment Guide
The message monitor is a standalone service: it is never embedded inside the gateway or the security and trust agent. Other components feed it tracked messages and notifications over the network, and it independently decides when a delivery notification sequence has gone unanswered long enough to warrant a failure notification back to the original sender's edge client.
There are two deployment models for the monitor. The cloud native model is the only model supported going forward and is where all new work happens; the legacy model is documented for HISPs still running an older release and planning a migration.
Cloud Native Model (recommended)
The monitor runs as a self-contained Spring Boot micro-service — direct-msg-monitor-sboot-<version>.jar — alongside the other reference implementation micro-services, communicating over REST and a message broker (RabbitMQ by default). It is a plain Java process with no application server to install, so it deploys unchanged onto bare metal, Docker, Kubernetes, Cloud Foundry, or a managed runtime such as Google Cloud Run.
This is the model used by the Cloud Native HISP deployment of the BareMetal Assembly Project, which shows the monitor in the context of the full HISP topology.
- Cloud Native Deployment — how to deploy and wire up the monitor micro-service.
Legacy Model
The monitor is packaged as direct-msg-monitor-web-<version>.war and deployed into a J2EE servlet container — in practice Apache Tomcat, alongside the configuration service and XD in the pre-9.0 BareMetal assembly.
The war artifact was last released at version 8.0.0 and is not published for 9.0 or later. The reference implementation retired Tomcat as of the 9.0 release. If you are running this model, treat it as a migration source rather than a deployment target.
- Legacy Deployment — configuration and operation of the war deployment, plus what changes when migrating.
Shared References
The following apply to both models:
- Configuration Reference — every configuration property the monitor reads, grouped by concern.
- Deployment Considerations — scaling, high availability, and shared state.
