Ask HN: What project/team-based developer metrics do you track?
Metrics such developer frequency, mean time to restore service, etc.
Curious to hear of other metrics being used and how they've proven useful.
Cheers! We track dev time spent typing, but only for gamification and individual productivity, using https://wakatime.com. Haha neat tool. Thanks! Cycle time (how long to complete tasks on average), number of customer found defects, MTTR (mean time to resolution) for production issues, SLO violations, velocity (story points completed on average), number of alerts, and probably more. We track it for ourselves, not for comparison with other teams. And these metrics are at the team level, never individual. Are you sure? At my company, managers say they don't measure and compare people but if you have a friend in management they will tell you the opposite. fwiw, one of our core principals is around being transparent where we can. If there are shenanigans going on, those are very unwelcome, especially the higher up you go in the org. Do some managers try to get individual metrics? I'm sure it is possible. But when managers have done similar side stepping of the intent of our values, the hammer comes down pretty hard on them. When it comes to reviews and such, it is more predicated on how well you've supported business goals, how well the team did so too, and some peer feedback. Managers are working with us daily, so it would be odd for the manager to not have direct knowledge on what the individual contributors do well or areas that are in need of improvement. That's nice. I hope you're right. Our company says the same things you are about business results and transparency, but they don't follow it behind closed doors. Planned vs. Unplanned work
Fairly easy-to-capture. Can be very coarse-grained and still have lots of meaning. (note: get this one ironed out before trying to calculate a Velocity metric) SLA - time taken per team to resolve tickets filed by users/internal staff Downtime of services Number of crash/non-fatal logs seen in logging services (you could be using Sentry, Firebase, etc) Just a few things tried and monitored at my company.