Open mcortesi opened 4 days ago
So we treat sources and fallbacks the same when writing? So I guess they are only differentiated by their read sequence logic? This looks fine to me, though I do have a strong preference for the config approach (KISS).
Alternative, on configuration processing check that a source can't be at same time cache and fallback
I'll let @epociask review this though given he wrote this feature.
Hi @mcortesi - there's already a config invariant to ensure that this case can never be true: https://github.com/Layr-Labs/eigenda-proxy/blob/main/server/config.go#L115-L120
@epociask amazin, haven't seen that.. maybe a comment on the code will help for the unaware reader
handleRedundantWrites
will do a write for each source incaches
andfallbacks
.While probably doesn't make much sense to have the same type of store be
cache
andfallback
they can technically be the same.So maybe check that
caches ++ fallbacks
is an unique set.Alternative, on configuration processing check that a source can't be at same time
cache
andfallback