Stream error in the HTTP/2 framing layer

### Expected Behavior

That a valid JSON response is returned
The service that I am using allows for passing oData query via query param.
This service works fine from Postman. My service provider downloaded Insomnia and and tried, he received the same error.

### Actual Behavior

Stream error in the HTTP/2 framing layer

A POST request with a payload & NO oData query param works fine
A POST request with a payload (that I know won't match any records) + an oData query param returns 204, which is fine
A POST request with a payload (that I know matches records) + an oData query param fails.
A GET request (different endpoints) + an oData query param works

If I look at the timeline, it appears that the service provider is marking the request as a 200 and that the issue occurs during the stream of the response. An endless stream of chunks is being returned to me before I get "HTTP/2 stream 0 was not closed cleanly". The response is not that big ... a single chunk .. not sure why it's streaming so much ...

Switching from HTTP/2 to HTTP/1.1 doesn't not fix the issue

Service provider thinks it has to do with URL escaping? But I am not so sure since the GET command works fine.

### Reproduction Steps

_No response_

### Is there an existing issue for this?

- [X] I have searched the [issue tracker](https://www.github.com/Kong/insomnia/issues) for this problem.

### Additional Information

_No response_

### Insomnia Version

2023.1.0

### What operating system are you using?

Windows

### Operating System Version

Windows 10 - 19044.2604

### Installation method

Insomnia updates itself

### Last Known Working Insomnia version

Never tried before this version

The payload names no source file, test, or reproduction steps. Start by reproducing the POST request with a matching-record oData query against the service, compare HTTP/2 and HTTP/1.1 behavior, and inspect the response stream; done means a valid JSON response is returned without the framing error.

Written by the indexing model from the issue text.