I'm working on a project using Django version 3.0.8 and PyMongo version 3.11.3. Currently, for each incoming request, a new MongoClient
is created, and the connection is destroyed once the request is processed. For example:
def get_all_logs(request):
req = json.loads(request.body.decode('utf-8'))
record_id = str(req.get('record_id')).strip()
with NewMongoClient(db='space').db() as db:
log_info_with_record_id = list(db.database_alternative.find({'recordID': record_id}))
Here, NewMongoClient
is a wrapper class around pymongo.MongoClient
I think this approach incurs too much overhead due to the constant creation and destruction of connections. Is there a more efficient way to reuse MongoClient connections across requests?
Yea, creating new MongoClient
for each request can be quite resource-intensive.
You could use Singleton solution for creating MongoClient
:
# get_mongo_client.py
from pymongo import NewMongoClient
mongo_client = None
def get_mongo_client():
global mongo_client
if mongo_client is None:
mongo_client = NewMongoClient(db='space')
return mongo_client
After this, you could use this in your view.py
:
# views.py
from .mongo_client import get_mongo_client
def my_view(request):
client = get_mongo_client()
with client.db() as db:
...
Thread-safe solution:
from pymongo import MongoClient
import threading
mongo_client = None
mongo_client_lock = threading.Lock()
def get_mongo_client():
global mongo_client
if mongo_client is None:
with mongo_client_lock:
if mongo_client is None:
mongo_client = MongoClient('mongodb://localhost:123/')
return mongo_client