Possible Memory leak

[jwoertink]

We have been using v0.4.0 in production since the release. When we try to update to the latest main commit, our app tanks.

These are graphs from our app showing memory use (top) and CPU use (bottom). The crazy spikes is when we deployed main , then after reverting back to v0.4.0 you can see where the app calmed down. We made no other changes for this deploy.

https://github.com/graphql-crystal/graphql/issues/48 is also somewhat related. Originally our app was getting DB::ConnectionRefused errors locally when pulling in the latest commit, but we made changes to our app which seemed to resolve that issue by just adding max_pool_size=170 to our postgres connection url.

I'm guessing on v0.4.0 there may also be a leak, then with https://github.com/graphql-crystal/graphql/commit/d43a6a129878fbe29068732169b6bca770da08b8 this was exposed a lot more by sending the calls in concurrently.