SSL connection has been closed unexpectedly

Hello everyone,
We have hosted our airbyte exports on bit.io postgres and we are running some dbt flows on them so that we can make any transformations needed in order to serve our Dashboards.
The thing is that when there was some development going on from out analyst, every month a NOTICE was coming up telling us that we were approaching our monthly limit of 1B rows. We were well aware of that it was not going to exceed the limit, although the notices made most of our query series to fail with a closed connection error.
This error stopped appearing when we upgraded plan and we were now paying per row scanned after the first Billion rows, and we never paid anything more, because we never exceeded the first Billion.
It was clear that the SSL errow was correlating with the limit, but deliberately producing some SSL connection errors did not limit the harm on our side. We run a very specific dbt flow every day with 67 sequential queries (and tests) and my bumping into SSL errors, could only make us trigger the flow from the beginning, leading us faster towards the limit.

The problem is that we are now approaching the disk space limit, knowing that we add only a couple of thousand rows more everyday, on our database, this creates a notification, which in its turn leads to more SSL errors, which leads to retriggers of the flow, which leads to even more queris, and thus it will at some point incur costs for us, as it will be sure that we will exceed the Billion.

Is this a known issue, is, bit.io trying to save costs of its customers, by making their job harder?
What is going on?

Welcome to bit.io and thanks for posting!

We were well aware of that it was not going to exceed the limit, although the notices made most of our query series to fail with a closed connection error.

Could you elaborate on this? Are the NOTICE messages being returned by bit.io breaking your postgres client (airbyte exports)? That is unexpected since NOTICE messages should not cause a postgres client to panic/close. How exactly are you connecting to bit.io?

It was clear that the SSL errow was correlating with the limit, but deliberately producing some SSL connection errors did not limit the harm on our side.

This is also surprising. Could you elaborate on how you were able to deliberately reproduce connection SSL errors? SSL errors often just indicate the connection was disconnected and the client did not gracefully handle it.

Is this a known issue, is, bit.io trying to save costs of its customers, by making their job harder?
What is going on?

We at bit.io are absolutely trying to make our customers lives easier! Sorry to hear that you are running into issues using bit.io. Rest assured that we are here to help investigate and resolve your issues.

Also, please kindly review The Connection Lifecycle. I suspect some of these issues that you are running into may be a part of the bit.io connection lifecycle (and the error messages you are seeing may not make it immediately obvious). Connection pooling (if feasible) may alleviate some of these disconnection issues.

Also, could you please share the entire error message(s) that your postgres client is seeing?