Serverless Data Streaming at Scale
The October virtual meetup tackled serverless data streaming at scale, drawing 127 attendees who were eager to learn about processing real-time data on AWS. This was an AWS Community Nordics event hosted by AWS User Group Copenhagen, continuing the cross-border collaboration that had become a hallmark of the virtual era. Anahit, a lead cloud software engineer at Solita and AWS Community Builder, delivered the talk, bringing hands-on production experience to the session.
Anahit covered the core AWS services for data streaming – Kinesis Data Streams, Kinesis Data Firehose, and Kinesis Analytics – along with how Lambda fits into the picture as a stream consumer. What made the talk especially compelling was the real-life example from Yle, Finland’s national public broadcasting company, where the architecture was streaming over 0.5 TB of user interaction data per day to enable near real-time content personalization.
One of the most valuable parts of the talk was the discussion around common pitfalls. Things like shard management in Kinesis, handling poison pill messages, and dealing with ordering guarantees are challenges that everyone hits eventually. Hearing about these from someone who had already worked through them at Yle’s scale saved attendees a lot of future headaches.
The serverless angle was particularly interesting. The idea of processing streaming data without managing any servers is appealing, but it comes with its own set of trade-offs around cold starts, concurrency limits, and error handling. Anahit was honest about where serverless streaming works well and where you might need to consider alternatives.
The 127 attendees made this one of the better-attended virtual meetups, and the engagement was strong. People were actively sharing their own streaming architectures in the chat and asking detailed technical questions. It was clear that data streaming was a hot topic in the Nordic cloud community, with many companies either already doing it or planning to start.
