Hi guys -
Mostly out of curiosity but, is it possible to run Kong with no cache, i.e. every node will hit the DB with every request?
Do you have any idea what kind of performance penalty would be suffered in that case?
In some cases (probably not ours, ha) I can imagine apps with rapidly-changing user data (the stuff typically stored in Kong’s DB for consumers and used/read by plugins to perform auth) and extremely stiff authorization requirements in which no margin of cache inaccuracy (auth is permitted because the request went through a not-yet-updated Kong node). Think something like a hedge fund with lots of employee turnover among their traders/execution folks…
Interested to hear what the community has to say on this topic since I’m quite sure it’s been discussed on many occasions and in many forms.