Just so we're clear, you're not using serverless mode correct? So what do you mean by "waking"?

My project is using Next.js + Postgres, and I have configured everything as serverless.

When I say “waking,” I mean: when a user accesses the Next.js app, and the app tries to connect to the database, the Postgres instance should wake from sleep and accept the connection — as it always did before.

However, since early yesterday morning (JST +9), many users reported they could not connect to the database. When I checked around 6:30 AM, the database was still stuck in sleep even though there was active traffic. The dashboard also showed the database in sleep state, not waking up.

This used to work fine before, and the behavior suddenly changed. Could you please help check this urgently?

Are you connecting via the database URL, or the database public URL

I am currently using the Database Public URL to connect.

Please use the private network

I will try it

Please let us know how it goes! Re-open this thread if it's still an issue

After our last conversation, I switched to using the private network (Database URL) for all connections. However, the same issue happened again — in the early morning of June 22nd (JST +9), the database did not wake up despite incoming traffic.

At this point, I’m no longer confident using the serverless mode in production, because the behavior is unreliable and has caused multiple user complaints.

Is there anything I can provide to help you investigate this further (e.g., logs, timestamps, request IDs)?

I really hope this can be resolved, as I would prefer to keep using serverless mode if it can be made stable.

❌ The internal ticket Serverless not waking up DB has been marked as canceled.

This is a small project, and every single user matters to me. Even a one-time failure to wake the database could mean losing a user that I may never get back. I sincerely hope your team can look into this as soon as possible.

Heya, our team will look into this, however we need more reports and reproduction steps before proceeding.

In the meantime, I suggest keeping serverless off on the database if it impacts you.

Of course — I’ve already turned off serverless on the database to avoid further issues.

Can your team access my logs to help investigate this?

Please let me know if there’s anything I should do on my end (e.g., enable specific logging, provide timestamps, or replicate certain traffic). I’m happy to cooperate to help you get more information.