search

eclipse-arrowhead / roadmap

Eclipse Public License 2.0
5 stars 9 forks source link

Common services like Echo and Monitor to be standardised across core system #38

Closed jerkerdelsing closed 2 months ago

jerkerdelsing commented 3 years ago

There starts to be a couple of services which show up at several core systems

The Echo service available in a couple of core systems seams to be slightly different documented and implemented in several of the core system.

The Monitor service is requested from application system by a couple of new and upcoming core support systems, e.g. PlantDescriptioEngine, QoS, ...

Can we device a common Echo service? probably YES. Can we provide this in as a library service?

Can we define a common Monitor service? Maybe, or can we find an approach where application system can have a library Monitor service which can be used by developer.

BlackRose01 commented 3 years ago

This could possibly be of interest: https://docs.spring.io/spring-metrics/docs/current/public/prometheus https://mydeveloperplanet.com/2021/03/03/how-to-monitor-a-spring-boot-app/

BlackRose01 commented 2 years ago

What should be part of the monitoring service(s)? Are we talking about used hardware/JVM ressources or active connections to the database or registered services (more statistic endpoints)?

jerkerdelsing commented 2 years ago

The monitoring service is still not defined to any details. The used HW/JVM resources may be addressed by mandating such info as meta data of Device, System and Service registries.

tsvetlin commented 2 years ago

What to monitor is implementation dependant. But Memory pressure is a generally a good indicatior to start with. This should be discussed in detail during the later WG Meetings.

About the Echo service, as it's current functionality it "only" returns a "Got it!" message, indicating the availability of the Core System.

Currently every core system behaves like this. It could be further improved, for example: instead of a string it could return a JSON object, with useful information about the core system, for example: name, version, etc.

jerkerdelsing commented 2 years ago

I will include the monitoring issue in the next Roadmap WG meeting agenda.

tsvetlin commented 2 years ago

Thank you!

BlackRose01 commented 2 years ago

Hello,

here is another solution for monitoring a java application (https://www.baeldung.com/spring-boot-actuators). Maybe it is better than metrics because it is more independend from prometheus and it runs on a management port (different web port, use other certificates, extendable).

KR, BlackRose

jerkerdelsing commented 2 years ago

@Per O to report on industry requirements at next Roadmap WG in Feb.

jerkerdelsing commented 2 years ago

@PerOlofsson to be addressed in the back box v.5.0 documentation

emanuelpalm commented 2 years ago

This issue has bearing on:

jerkerdelsing commented 2 months ago

A SysMonitor SysD and Monitor SD is available in v5.0 documentation. The issue is thereby closed.