ETOOBUSY 🚀 minimal blogging for the impatient
I did a bit of scouting but decided to roll my own.
After Thinking on a data collection API a bit, I ventured into the internet to find out ready-made things.
In recent years, Time-series databases have become a hot topic, and yet I feel they’re not really the right thing for my low-spec requirements: collect some data every now and then, for a handful of people.
Many of them, for example, are optimized for collecting data like crazy and support different levels of aggregation. This makes sense when there’s a lot of data coming, of course. On the other hand, I anticipate very sparse data, so I’m more worried about pre-allocating all the space over 50 years or so than about data possibly exploding in time.
Hence, for the time being I’ll be collecting and tracking individual data points in a traditional RDBMS with a simple API; in case scale will ever by an issue, I’ll figure out how to ingest all the collected data in a new system.