Using Farma as the backend for Pingem has provided good insights on how Farma should evolve:

1. Make the kv-store simpler. I was trying to do things that are hard to do with a kv-store -it worked, but the increased complexity is hard to manage.
2. Use NATS to push events to other services like Pingem, and let these services do the "smart" things like reporting and data mining.
3. Take advantage of NATS to implement clustering: Run more than one Farma instances both for load balancing and resilience.

The API will probably change (not much), so if you are using Farma in your project (or through the Pingem API), let me know.

Using Farma as the backend for Pingem has provided good insights on how Farma should evolve:

1. Make the kv-store simpler. I was trying to do things that are hard to do with a kv-store -it worked, but the increased complexity is hard to manage.
2. Use NATS to push events to other services like Pingem, and let these services do the "smart" things like reporting and data mining.
3. Take advantage of NATS to implement clustering: Run more than one Farma instances both for load balancing and resilience.

The API will probably change (not much), so if you are using Farma in your project (or through the Pingem API), let me know.

https://farma.vrypan.net/