Serverless – Introduction

Serverless is a very popular topic at the moment in the IT industry.

For monitoring (and hence I see this a lot recently in my daily job) it creates challenges as one cannot longer instrument serverless functions with traditional agent-based technologies. AWS is pushing it from their side, they are creating a lot of noise about it at the AWS summits this year, which I also mentioned in a previous post. Additionally, AWS pricing seems to be made to attract customers (first 1 million invocations are free regardless of free-tier or not). A Cloud Guru, who I used for my AWS exam, claim to run everything on serverless, they even are hosting a conference dedicated to serverless for 2 years now.

I had developed features in Harvee which never went into production. Mainly due to resource costs when deployed on the Internet but also because of recent changes in data privacy regulations as some of these features deal with creating a graph of information around events that people go to or talk about (notifying friends about events or downloading event details into one’s calendar would be stored in the graph). However, there is this one feature which can be purely driven by the data that I collect from the event source and correlating event information and that is to determine events in venues in the proximity of a selected event. So this I decided was a good first use case for a serverless function and I decided to implement it on AWS Lambda.

Roughly two things had to happen for this to become reality: I had to adjust infrastructure (run a graph database, populate it with nodes and relationships regarding events and venues) and write the AWS Lambda function to query events in proximity. There will be two subsequent posts coming soon covering these two set of activities.