The production service is the service containing the ML model.
It’s important to implement serious ML monitoring, as we need to make sure that the service itself works fine.
To monitor the service health you can measure some metrics like up-time, memory-usage, latency and others.
very important metrics
These metrics are critical, if you can only measure few metrics then it should be the service health metrics.