Kubernetes Operator for scaling in response to inter-service back-pressure for reactive systems

[Gordon Hutchison]

Hi MicroProfilers,

With the MicroProfile Reactive Messaging Specification handling reactive back pressure ticketing automatically

on behalf of applications...and this being a common book-ending code for more than one 'Connector' technology being used between

services {Kafka, REST etc...}.

Would it be possible to have a Kubernetes operator function that scaled in response to the abundance or scarcity of

reactive back-pressure request tickets.

It might help push MicroProfile more into the 'natively cloud native' space and more into reactive systems from

reactive programming.

Gordon

[Gordon Hutchison]

Or perhaps, I learn (thanks to Marc Sluiter)

   https://kubernetes.io/docs/tasks/run-application/horizontal-pod-autoscale/

with appropriate custom metrics:

   https://github.com/kubernetes/community/blob/master/contributors/design-proposals/instrumentation/custom-metrics-api.md

based on back pressure status

There is already:

   https://github.com/eclipse/microprofile-reactive-messaging/issues/31

for Reactive Messaging Metrics from Andrew Rouse (who also did the metrics support in Fault Tolerance)

So possibly part of that chunk of work.

Gordon.