Things about serverless I wish I used from the start

Antonio Terreno
Feb 22, 2019 · 2 min read

This list comes from the time ‘wasted’ going through multiple git repositories, changing, pushing, one by one all the serverless.yml file, making ‘horizontal change’ across all functions: if I knew these things existed when I started, I would have saved myself quite some time


You can use SSM Parameter Store and it’s dead easy to use, no more hardcoded variables or secrets, those variables get expanded at build time by Serverless, we started to use them also for not-so-secret stuff, it’s nice to have a versioned, auditable repository with all those values in AWS.

versionFunctions: false

After a few months, we ran out of space on the Function and layer storage, which is by default 75 GB. I ended up writing a Janitor lambda function to remove all the non-live lambda deployments, but I much rather work with a always roll forward policy, so all our functions now are basically not versioned, something less to worry about.

logRetentionInDays: 30

Another parameter that you probably want to specify is the logging retention policy, this will avoid over cluttering (and overpaying) your Cloudwatch bill. 30 days is still probably too much, as we tend to fix whatever is wrong in less than 24 hours.

name: your-serverless-lambdas-deployments-bucket

Without setting this parameter things get a bit out of control with randomly generated buckets names, with this, all the deployments will end up in S3 in a nice and tidy structure such as serverless/{service-name}/{stage}.

- serverless-log-forwarding

We recently introduced centralised logging, Cloudwatch is far from being usable as logging system, we use this plugin and forward our logs to a lambda that forwards then the logs to Papertrail.

iopipeToken: ${file(${opt:stage}.yml):iopipeToken}

I always mention IOPipe and we used this since the very early days, but what I didn’t realise until recently is that the free plan offers two projects, so you can actually split your monitoring in prod and non-prod environments and set up what they call ‘global’ alerts: no need to set up a new alert every time you introduce a new lambda.

A last thought: we implemented our lambdas in multiple git repositories, every repository might have from one to fourteen functions each, but what if we were doing a mono repo? We could have one single serverless.yml file, with the common configuration and then many other yml files containing function specific configuration. I think it might work really well, the serverless framework development moves really fast and to fully leverage the plugins system I think it would be easier to embrace a mono-repo structure.

I’d love to hear what are your favourite, must-have plugins and some war stories on serverless and mono-repos.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch

Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore

Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store