Under constant load, backend refresh doesn't provide efficient mechanism to save enought backend request about cache-manager

[jaredwray]

@lchenay - that is something that could be possible but I would think that we remove the background refresh at scale as that causes other issues or go to a Kafka like infrastructure to update the caching stores directly which is what most larger scale enterprises do. Putting this into the library could cause multiple issues when you have 100s of nodes trying to do a refresh.

[jaredwray]

@lchenay - let me know if you want to do a video call on it and I can show you what we are doing with this type of architecture.

[lchenay]

I would love to switch to a read model infrastructure. Sadly it's not easily possible and we have to deal with it in short term.

Totally agree on your feedback and the decision to not put it natively in the library.

We will find a solution on our side