We have used serverless cloud infrastructure for the last five years and we have been spoilt, just how spoilt was reinforced lately

#AgileData

When we started building AgileData five years ago we made the decision that one of our core architectural principles would be the use of serverless infrastructure components by default.

This has been invaluable over the years, we only pay for what we use, it allows us to scale when we need to and removes the cost of creating, deploying and managing containers etc.

Every time we need to fill a gap in our solution we look at the serverless options available in Google Cloud and decide if we can use one to fill this gap. But every now and again there is something where a Google Cloud capability is not available.

There is a capability in the AgileData Platform that we have needed to add for a while, and no Google Cloud Serverless option exists. It’s time we filled that gap so we looked further a field to see if there was something we could lease, or if we need to build something ourselves.

We found something that would do the job, and while it wasn’t a serverless option, it was based on a container deployment which we could multi-tenant, so we could make it work.

While it wasn’t cheap, it was in the same cost realm of other tools that did that data work.

But the killer for us, was we had to commit to a certain number of licences, even if we used them or not. And we had to commit to those licences for 12 months in advanced.

And that’s when I realised how spoiled we have been using Google Cloud serverless capabilities. It really has allowed us to build our product and our company in a way that would not have been possible 10 years ago.

And in case you are wondering, we are going to build this capability, its time to fill that gap.