Support Metrics
kyuff opened this issue · comments
PO aspires to be the lower tier of an application stack by replacing the Data Access.
There is a need to follow Metrics for that layer and PO should support that.
Introduce an interface that observes relevant Metrics. It should have out of the box implementations for Prometheus and be easy to implement for other monitoring systems.
Some examples of relevant metrics:
- Delay between storing a message and assigning it's group number
- Delay for Subscribers between a stream and how far it is
- Rate of messages per Group/Stream
Some work have been prepared for this with #33, but it still requires attention to the three listed abilities.