Cloud Firestore seems to get stuck in Python when multiple concurrent queries are sent right after the app intialisation

Question

I am trying to retrieve some sets of documents from Cloud Firestore, each corresponding to a different query. In order to speed up this process, I am using Python's library concurrent.futures, so that those queries are performed concurrently. The thing is that I've come across a weird behaviour: if those queries are sent right after the app object is initialised, Firestore seems to get sort of stuck, and no response is returned ever (or at least in a reasonable time), as if the app wasn't fully initialised and couldn't handle a stack of queries after. My solution consist in performing a dummy query in between the initialisation and the concurrent queries, and my doubt is simple: is this behaviour expected or am I missing something?

This is the code that I'm using, where lines 15 to 17 sohuldn't be needed.

import concurrent.futures
from firebase_admin import firestore, initialize_app, credentials

print("Initializing app session with credentials")
credential_json = '/path/to/your/credentials.json'
firebase_credential = credentials.Certificate(credential_json)
app = initialize_app(name="your-app-name",
                     credential=firebase_credential,
                     options={"projectId": "your-project-id"})

db = firestore.client(app)

print("App initialized")

# This line is necessary to prevent concurrent processing from crashing but it shouldn't be needed to fully initialize the app
db.collection("a").document("b").get()
print("Dummy query")


# Function to fetch a single document
def fetch_document(doc_ref):
    print(f"Fetching document {doc_ref.id}")
    doc = doc_ref.get()
    print(f"Fetched document {doc_ref.id}")
    return doc.to_dict()


# Function to fetch multiple documents concurrently
def fetch_documents(document_refs):
    print("Initializing thread pool")
    with concurrent.futures.ThreadPoolExecutor() as executor:
        # Submitting tasks for fetching documents concurrently
        print("Processing document list")
        future_to_doc = {executor.submit(fetch_document, doc_ref): doc_ref for doc_ref in document_refs}
        # Collecting results
        results = {}
        for future in concurrent.futures.as_completed(future_to_doc):
            print("Document/job processed")
            doc_ref = future_to_doc[future]
            try:
                print(f"Saving processing results for {doc_ref.id}")
                results[doc_ref.id] = future.result()
            except Exception as exc:
                print(f"Error in document {doc_ref.id}: {exc}")
                results[doc_ref.id] = str(exc)  # Record any exception that occurred
        print("Document list processing complete")
    print("Closing thread pool")
    return results


doc_ids = ["doc_id_1", "doc_id_2", "doc_id_3"]
# List of document references
refs = [db.collection('collection_name').document(doc_id) for doc_id in doc_ids]

# Printing results
for doc_id, data in fetch_documents(refs).items():
    if isinstance(data, dict):
        print(f"Document {doc_id}: {data}")
    else:
        print(f"Error fetching document {doc_id}: {data}")