Hi all,
I wanted to share some testing insights and invite suggestions on optimizing resource usage for OpenRemote deployments.
After experimenting with a few configurations, I’ve manage to deployed the V1.5.0 platform on an AWS Graviton (ARM-based) EC2 instance with 8 vCPUs and 32Gb memory.
Initially, I ran into performance issues on smaller x64 and ARM CPU instances — particularly with the Manager container failing and becoming unhealthy, and rule compilation seeming to be very slow. Upgrading to the 8-core instance resolved those issues, and the system has been spinning up a lot faster and running responsively under moderate telemetry and rule loads.
That said, the start up process still feels sluggish compared to a test I ran on my Windows laptop, where everything spun up extremely quickly. Given that we’re not managing a large number of assets or complex rule sets, relying on a more powerful (and more costly) AWS instance isn’t ideal.
I’m keen to know:
- Has anyone applied specific JVM tuning, Docker memory limits, or PostgreSQL optimizations to reduce the system’s footprint?
- Any tips for improving start up time or avoiding container health check issues?
- etc
I am using the docker-compose.yml set up from the Quick Start guide modified for the machine’s IP address and Keycloak credentials etc.
Thanks in advance
Cheers,
Clint