How do I use Knex with AWS Lambda?

I’m running into issues with connection pooling while testing some code. I’m expecting my lambda function to be called maybe several thousand times over a couple of seconds, and having trouble figuring out the best way to connect to my db. Here’s a very similar issue for node/postgres: essentially the problem is that I need to be able to get a connection from the pool if one is available, however I can’t depend on the pool existing because of how AWS (unreliably) reuses lambda containers.

Basically what I’m looking for is a way to reliably get or create a connection to my db. I have not been able to find any examples (like while(!availableConnections) { tryToGetConnection() }. Do I need to interact with node-pool? How can I do this with Knex?

[elhigu]

Sorry I committed in a middle of sentence, my kid just threw like 3 liters of water to floor 🥇 I’ll update the comment above in a moment…

[jamesdixon]

I’ve had Knex running in Lambda for almost a year now with no problems. I’m declaring my Knex instance outside of my Lambda function and utilizing context.callbackWaitsForEmptyEventLoop = false as mentioned in other posts.

That said, over the past day, something seems to have changed on the Lambda side as I’m now seeing a huge connection spike in Postgres; connections don’t seem to be closed.

Has anyone else using the aforementioned approach seen any chances over the past day or so?