Deployed to cloudflare workers with supabase gives 15-20s connect time. What gives?

[daneatmastra]

This issue was created from Discord post 1493589587619086418:

Worked well in a container but that was too slow to start so i moved to workers. Now the app itself boots super fast but the connect to supabase averages 15-20 seconds. Can't for the world figure out why. I'm creating one mastra db in supabase to handle multi-tenancy so i can't use hyperdrive efficiently (if that would make a difference)

[daneatmastra]

Thank you for reporting this issue! We have labeled it for trio-wp and we will look into it as soon as possible.

If you're experiencing an error, please provide a minimal reproducible example whenever possible to help us resolve it quickly.

[mikhael28]

I have responded to the user on Discord, waiting for their response; asked for a reproducible sample and gave a few pointers on connection pooling with Supabase in case it was related.

Hi @js - would you please send a reproducible example of your code to the GitHub issue created above? We aren't able to really do an investigation without it. A few things to consider as well on your end; if you are using a free Supabase subscription, the first db connection after a lull/pause takes 10-20 seconds. Could be a part of the issue. Second, are you pooling your connections properly? Cloudflare workers are stateless and ephemeral; they aren't capable of maintaining a persistent TCP connection, so each connection to connect to Postgres is expensive. Use the built-in Supabase connection pooler (PgBouncer) on port 6543. There are a couple of other things you can check on your end, but I think those are the two most likely culprits outside of Mastra - like I said, please send a reproducible codebase sample to us for a proper investigation.