Monitoring isn't just for ops people and application developers. Front-end developers need just as much visibility into their work as anyone else. How is the front-end developer's relationship to production formed?
When thinking about the things we monitor with our configuration-managed monitoring pipelines, we rarely include the front-end development team or consider their needs from the outset. As the statistics that browser APIs can collect have become more robust, it is becoming easier to standardize the collection of WebPerf stats.
Inspired by the UK GDS team's event-store project, the team at Critical Mass created a Go program that can collects Navigation Timing API data, Javascript Error Reports, and CSP Reports. The application became the focus of one of our monitoring recipes, and has since become standard on any node serving web pages to end-users. Our front-end developers have developed drop-in javascript functions to interact with this service and thus made WebPerf monitoring relatively turnkey.
In this ignite, I talk about the conversations and brainstorming that lead to this initiative, about some of the mechanics of collecting the data, and how it has changed the way front-end developers relate to the production applications and monitoring dashboards used by the team. I focus on the cultural aspects of the experience and outcomes, but the code and architecture is discussed here: http://blog.lanyonm.org/articles/2015/03/29/golang-http-stats-collector.html.