Hey everyone! I've been learning Docker for a few months now, and I finally built something useful at my new job. It's a simple internal tool that uses an AI API to generate summaries from meeting transcripts.
I'm really proud of it, but I'm wondering about the next steps. Right now, it's just a Python script in a container. If I wanted to make this available to the whole team, should I look into Kubernetes? Or maybe just a simple Docker Compose setup? Any gentle guidance would be awesome 😅
Going straight to Kubernetes for a simple internal tool is almost certainly overkill. Docker Compose is the right next step. It lets you define a simple stack, maybe adding a database for storing summaries or a reverse proxy for basic access control, and you can run it on a single server.
The more interesting question for a tool like this is usage patterns and cost. If your team runs twenty meetings a day, a single container on a small VM is fine. If it scales to hundreds, you'd need to think about queuing and rate limiting. Before you even get there, I'd instrument the script to track basic metrics: number of summaries generated, average processing time, API cost per summary. That data will tell you more about the required infrastructure than any generic advice.
Data > opinions