If installing on kubernetes, traefik is required as ingress controller. You can run both nginx and traefik side by side, but you cannot have nginx infront of traefik. A storage provider needs to be provisioned that support both up and down scaling the size. Using local storage and assigning pods to specific machines is NOT supported or recommended. If you need to share RabbitMQ with other applications, its recommended create a virtual server solely for OpenFlow. In the connection string then add the name of your virtual server.
Using OpenFlow without premium features:
using docker, with traefik as ingress controller allocated around 200 to 300mb ram for RabbitMQ allocated around 200 to 300mb ram for each API node allocated around 100 to 200mb ram for traefik allocated around 120 to 250mb memory for each NodeRED you want to start allocated around as much ram as possible for MongoDB, but at least 200mb in total for a minimum setup 1Gigabyte of ram each image is around 500 to 1 Gigabyte and most setups takes a long time to reach the first 1 gigabyte db storage milestone so allocated at least 10 Gigabyte of disk space
1) for option to use Grafana toward OpenFlow data This requires only starting a Grafana instance and should not require more than 50mb to 100mb of RAM (the image is 250mb so also 500mb of disk space ) 2) option to use Open Telemetry to collect usage, metrics and spans and send custom tracing info from NodeRED. There are a few options here, but a typical setup would involve: allocate around 200mb to 1 gb memory for otel-collector allocate around 300mb to 3 gb memory for victoriametrics allocate around 200mb to 500mb memory for jaeger allocate around 500mb to 3 gb memory for cassandra, and enough disk space to hold 14 days of metrics and spans. ( 50 to 100 Gi disk space is a good starting point )