Moongose ready state and aws lambda - cache connection - mongodb

Question

I am trying to implement a connect to database function that uses mongoose (MongoDB), and AWS lambda functions. Best practice, as I understand it, is to cache the database so that the connection can be reused. This, I have working. However, the problem is that the mongoose.readyState variable uses some fixed amount of time to determine if the connection is valid or not.
I was wondering if there is an alternative to .readyState to check if the connection is valid. I know that I can try to run a random query against the database every time I want to reuse the connection. Thus, I could determine if an exception happens or not. However, this feels a little hackish.
Here is my connect to database function:
let cacheDb = null
let options = {
useMongoClient: true,
autoIndex: true,
autoReconnect: true,
keepAlive: true,
socketOptions: {
keepAlive: true,
autoReconnect: true,
connectTimeoutMS: 30000
},
reconnectTries: Number.MAX_VALUE,
reconnectInterval: 200,
poolSize: 1,
bufferMaxEntries: 0,
};
mongoose.Promise = global.Promise
module.exports = {
connectToDatabase: function connectToDatabase(context) {
context.callbackWaitsForEmptyEventLoop = false
if (cacheDb && mongoose.connection.readyState == 1) {
return cacheDb
}
else {
mongoose.connect(process.env['MONGODB_URI'], options, function(error){
if(error){
console.log(error.toString())
}
cacheDb = mongoose.connection
return cacheDb
})
}
}
}
Do you have any idea as to an alternative approach?

Answer 1

In order to check for a valid connection , you should be checking your cached db serverConfig.isConnected method.
like so:
if (cachedDb && cachedDb.serverConfig.isConnected()) { ...
This is according to mongodb, in this helpful article:
https://www.mongodb.com/blog/post/optimizing-aws-lambda-performance-with-mongodb-atlas-and-nodejs