Slow message write to database in production - tested with upstash/redis and firestore

[myhendry]

anyone come across a situation in production, where it is very slow to write to a persistent store eg upstash/redis or firestore using langchain's BufferMemory? it streams nicely and i'm able to output the response and its able to have short term memory of the conversation. however, it is very slow and inconsistent when writing to a persistent database in production. it works fine in development. I tested using upstash/redis and firestore as store and i get the same issue. I also switched to using runtime="nodejs" instead of "edge". ll appreciate if anyone has encountered same issue and how you go about solving it. thanks