Got timeout error -1001 after some HTTP requests using Swift - swift

Question

I create an http request using PUT to get data from the server. I got this data from the server and transform it on a PDF file.
When I run it for the first time everything runs fine, but after some calls, I start to get timeout errors.
Sometimes, I need to restart the app to be able to receive HTTP requests again.
This is the code that I use.
func callGetPdfFromEndpointUsingNSMutableURLRequest() {
if codigoBarra == "" {
messageError = "Código não localizado"
showingAlert = true
redirectToPdfView = false
showingActivityIndicator = false
return
}
let serviceRepository = ServiceRepository()
// let codigo_barra = "d152d36914313fedfbf36842a7195b723"
let json: [String: Any] = ["codigoBarra":"\(codigoBarra)"]
let request: NSMutableURLRequest = serviceRepository.clientURLRequest(endpointPesquisa, typeAutho: .basic, parms: "", body: json as Dictionary<String, AnyObject>)
print("request: \(request)")
print("request.url: \(String(describing: request.url))")
serviceRepository.put(request, retryLogin: true, completion: {isOk,msgError,httpCode,needLogin, response in
if isOk {
tratarRequisicaoPdf(response)
} else {
print("erro no request - is not ok | - httpCode: \(httpCode)")
var stringResponse:String = ""
if response != nil {
stringResponse = String(data: response as! Data, encoding: .utf8)!
} else {
stringResponse = "Sem resposta do servidor, tempo limite da solicitação foi esgotado."
}
messageError = "\(stringResponse)"
print(messageError)
showingAlert = true
redirectToPdfView = false
semaphore.signal()
}
semaphore.wait()
showingActivityIndicator = false
})
}
This error is unstable, sometimes it shows, sometimes it don't appear.
The people working on backend was not able to detect any problems.
I got the following error:
2022-05-20 15:33:15.442419-0300 CDA[2016:38068] Task <147B6F7F-E46A-47D0-A258-D6F3E5417D7E>.<1> finished with error [-1001] Error Domain=NSURLErrorDomain Code=-1001 "Esgotou-se o tempo limite da solicitação." UserInfo={_kCFStreamErrorCodeKey=-2102, NSUnderlyingError=0x7fd0e5245520 {Error Domain=kCFErrorDomainCFNetwork Code=-1001 "(null)" UserInfo={_kCFStreamErrorCodeKey=-2102, _kCFStreamErrorDomainKey=4}}, _NSURLErrorFailingURLSessionTaskErrorKey=LocalDataTask <147B6F7F-E46A-47D0-A258-D6F3E5417D7E>.<1>, _NSURLErrorRelatedURLSessionTaskErrorKey=(
"LocalDataTask <147B6F7F-E46A-47D0-A258-D6F3E5417D7E>.<1>"
), NSLocalizedDescription=Esgotou-se o tempo limite da solicitação., NSErrorFailingURLStringKey=https://myurl.sp.gov.br/gedave/api/spservicos/v1/buscaRequisicaoExame, NSErrorFailingURLKey=https://myurl.sp.gov.br/gedave/api/spservicos/v1/buscaRequisicaoExame, _kCFStreamErrorDomainKey=4}
response::: nil
erro: response is nil
httpCode: 0
What can I do to try to detect what was causing the timeout errors?
Edit:
I added the source code that can be viewed here https://swiftfiddle.com/mz2dxw6z6bda7a6t44cryncrpi.
NEW EDIT:
Answers to comments created by #Rob
Good to know about the NSUrlMutableRequest
I will try to use the 'finishAndInvalidate' in my URLSession. I didn't know about that.
My problem is unpredictable. Now I start the app and the first call got an timeout, after the second call the app works. Sometimes it starts working, but after some requests, I got a timeout

Answer 1

In the absence of a MCVE, there is not enough here to diagnose or reproduce the problem.
That having been said, there are some interesting clues:
You are calling wait (on a semaphore). Eliminate the DispatchSemaphore and if you want to know when the requests are done, use DispatchGroup. But when you use the dispatch group, use notify when it is done, not wait. Avoid blocking threads unnecessarily. And never block the main thread.
Your network request is performed by ServiceRepository, which you have not shared with us. But in your fiddle link, you show us unrelated URLSession code that is specifying the main queue as its delegate queue. If ServiceRepository is doing something similar, that, combined with the wait, above, could easily deadlock your code.
So, eliminate the semaphore, avoid ever calling wait (whether semaphore or dispatch group), and the deadlock risk is eliminated.
That having been said, that is only one potential timeout risk. The other scenario might be that you are simply issuing too many requests for the URLSession to run them without timing out.
If that is the case, you have a few options:
Increase the timeout threshold (either of the request or the session).
Try bumping up the timeout values and see if that mitigates the issue.
Submit uploads in just-in-time manner.
E.g., you might have a process where, rather than initiating all the uploads at once, that you issue each upload upon the completion of the prior upload. Or you can wrap the upload in a custom, asynchronous, Operation subclass (which is complicated in its own right), and then tell the operation queue to only attempt 4 at a time, or whatever. There are many techniques to tackle this, but the idea is to prevent timeouts by not even submitting the upload until you are confident that the URLSession can perform them.
Use background URLSession with file-based uploads.
If you have uploads that are so slow that they risk timing out, it begs the question of whether you really want to force the user to keep the app running in order to let the uploads finish at all. If you use a background URLSession, the background daemon will not let them timeout, but rather will upload them when it can.
Refactoring this for background URLSession is a non-trivial exercise (the first time you do it, anyway). You have to abandon completion handlers, use delegate-methods API, use file-based uploads, etc. But it is an incredibly powerful way to let uploads proceed in the background, allowing the user to leave the app, etc.

Answer 2

I don't see full code but you should have a look into semaphore usage - probably wrong multithread logic leads to not complete you request and your completion is hovered for a long period and causes URLDataTask to produce the timeout error