MongoDB Atlas Timeout After Deployment Update — "ReplicaSetNoPrimary"

Question

I updated my backend deployment (Flask app) with a new feature not related to database logic, and suddenly MongoDB Atlas connections no longer work. Before the update, everything was fine.

Local environment still works → only deployed instance fails.

---

✅ App startup (normal)

Environment loads correctly:

MONGODB_URI: mongodb+srv://...
DB_NAME: expTracker

App + scheduler boot normally, health checks pass.

---

❌ Error during DB call

Calling db.users.find_one() fails:

pymongo.errors.ServerSelectionTimeoutError:
... timed out ...
topology_type: ReplicaSetNoPrimary
server_type: Unknown
ac-xxx-shard-00-00.mongodb.net:27017 timed out
ac-xxx-shard-00-01.mongodb.net:27017 timed out
ac-xxx-shard-00-02.mongodb.net:27017 timed out

So no shard becomes primary; all time out.

No code or URI changes were made.

---

Code Snippets

MONGO_CONNECTION_ARGS = {
 "tls": True,
 "tlsDisableOCSPEndpointCheck": True, # This was in the original, keeping it
 "serverSelectionTimeoutMS": 8000,
 "connectTimeoutMS": 5000,
 "socketTimeoutMS": 10000,
}
def get_db():
 if 'db_client' not in g:
 uri = current_app.config['MONGODB_URI']
 # This MongoClient constructor fails in deployment
 g.db_client = MongoClient(uri, **MONGO_CONNECTION_ARGS)
 g.db = g.db_client[current_app.config['DB_NAME']]
 return g.db

What I’ve checked

✅ Same URI works locally
✅ Env vars printing correctly
✅ Service starts successfully
❌ Deployment instance cannot query MongoDB

• Joe: This looks like a connection failure - any additional detail in the error messages?
• SOEUNG Phearaneron: Just ServerSelectionTimeoutError with topology_type=ReplicaSetNoPrimary, and all shard hosts time out. also, i’m running on Koyeb. if I change the MongoDB user password, it works once, but after the next redeploy it fails again with the same timeout.
• Wernfried Domscheit: What is your connection string (without password)?