As time goes by, I begin to notice that my contributions to open source software are not groundbreaking improvements in performance, nor are they impressive new features. Mostly, it seems, I’m noticing small mistakes and fixing them. What’s more, most of those mistakes aren’t bugs in the code, but rather typos and other mistakes in documentation. It’s tempting to view pull requests for documentation as second-class citizens; after all, these changes don’t improve the application itself.

Wednesday morning I, along with operations folks all over the country, started getting alerts from systems that depend on Amazon’s S3 object storage. Quarqnet uses S3 to store uploaded data from athletes using our Qollector. We also depend on S3 to store some files that we occasionally read in order to provide our services. S3 is one of Amazon’s oldest and most reliable services, and a great many online systems depend on S3 to store data.