#458 – Site Reliability

This week, we met with a customer who had their meeting interrupted by a crash.

In our 10 years of operation, we have had only a handful of unplanned incidents on our server infrastructure that have affected end users. When an incident happens, we take it seriously. Meetings are expensive. The tool has to work. It’s stressful when things don’t work smoothly.

The Parabol application is evolving at a rapid pace. We’ve been busy preparing our Pages knowledge management and Insights features for general availability. Two weeks ago we introduced some changes that caused a server exception under rare conditions. Our application and infrastructure have layers of redundancy that normally would shield end users from this class of bug. In this case, however, one user had their retrospective become unresponsive.

Our monitoring systems noticed the failure immediately and we jumped into action to produce a fix. By that time, at least one team had a bad time.

We take incidents like these seriously. We should! After all, we’re makers of a popular incident post-mortem tool. Prevention, detection, reflection, and reaction are core practices in our culture.

So is listening. If ever you’ve had an experience that is less than perfect, please let us know. We’d love to hear so we can improve it for the future!

The week following the U.S. Labor Day holiday tends to mark the start of the period of time we call “back to school season” and an increase in platform activity. Even with the short week, we’ve seen an uptick in meetings ran. We’re eager to see if MAU will follow suit in the coming weeks.

…crafted pen and ink prototypes from our innovation advisory board meeting a couple weeks ago.

…onboarded new accounts to our Department of Defense hosting partner, Platform One.

…released Pages to nearly 400 organizations in our free tier. We’re looking to collect more data as to how these big changes perform in production (and if anybody writes in with any questions or feedback).

…fixed several bugs gating our release of Pages to general availability. As of this writing, just two more bugs to go and we’re ready to ship!

…we’re likely to release Pages to production.