MongoDB

Managed MongoDB with replica sets, automated backups, point-in-time recovery, and private cluster networking.

Creating a MongoDB instance

From the dashboard, click New service → Database → MongoDB. Choose MongoDB 7.x, a storage size (1 GB to 500 GB), and a plan. StackBlaze provisions a StatefulSet and initializes the replica set automatically.

Attach MongoDB to your service from the Environment tab. StackBlaze injects MONGODB_URL automatically.

Connection strings

Internal (recommended)

mongodb://user:password@[service-name].internal:27017/dbname

Replica set connection (Pro+)

On Pro and Enterprise plans, MongoDB runs as a 3-node replica set. Use the replica set connection string to enable automatic failover and read preference configuration.

mongodb://user:password@[service-name]-0.internal:27017,[service-name]-1.internal:27017,[service-name]-2.internal:27017/dbname?replicaSet=rs0&authSource=admin

External (for Compass / mongosh)

Use the external connection string for MongoDB Compass or mongosh from your local machine. Find it in Settings → Connection.

mongosh "mongodb+srv://user:password@cluster.stackblaze.cloud/dbname"

Replica sets

With a 3-node replica set, MongoDB automatically elects a new primary if the current primary becomes unavailable. Your application should use the replica set connection string and enable retryable writes for seamless failover.

import mongoose from 'mongoose'

await mongoose.connect(process.env.MONGODB_URL, {
  retryWrites: true,
  w: 'majority',
  readPreference: 'primaryPreferred',
})

Automated backups

All plans include daily automated snapshots using mongodump for Free/Starter and filesystem-level snapshots for Pro/Enterprise. Backups are encrypted with AES-256 and stored in a separate availability zone.

Point-in-time recovery (PITR)

Pro and Enterprise plans support PITR via continuous oplog tailing. StackBlaze ships oplog entries to object storage, enabling recovery to any point within the retention window.

Indexes

Create indexes in your application startup code or migration scripts. For production deployments, always create indexes with { background: true } (MongoDB < 4.2) or as a rolling build to avoid blocking reads and writes on large collections.

// In your model definition
const userSchema = new mongoose.Schema({
  email: { type: String, unique: true, index: true },
  createdAt: { type: Date, index: true },
  'profile.city': { type: String },
})

// Compound index
userSchema.index({ 'profile.city': 1, createdAt: -1 })

// Text search index
userSchema.index({ name: 'text', bio: 'text' })

Connecting with common ODMs

Mongoose

import mongoose from 'mongoose'

if (mongoose.connection.readyState === 0) {
  await mongoose.connect(process.env.MONGODB_URL!, {
    serverSelectionTimeoutMS: 5000,
    socketTimeoutMS: 45000,
  })
}

Prisma (MongoDB provider)

datasource db {
  provider = "mongodb"
  url      = env("MONGODB_URL")
}

model User {
  id    String @id @default(auto()) @map("_id") @db.ObjectId
  email String @unique
  name  String
}

Motor (Python async)

from motor.motor_asyncio import AsyncIOMotorClient
import os

client = AsyncIOMotorClient(os.environ["MONGODB_URL"])
db = client.get_default_database()