Buffer, as an organization, closes the final week of the yr yearly, and traditionally, as an engineering staff, we’d do a code freeze for the week earlier than we’re closed so nothing broke heading into the vacations. (Fairly sensible of us.)
This yr, we did one thing totally different, and it nonetheless labored out. We used the final week of the yr as a “fixathon” for our complete engineering staff.
Fixathon = every week the place our 29-person engineering staff spent an entire week targeted on fixing bugs.
By the tip of the week, not solely had we fastened 44 bugs, we additionally didn’t break something within the course of. 😎
Right here’s extra about our course of for establishing the Fixathon and the outcomes.
Why we spent every week on bug fixes
Like all product, there are bugs in Buffer. We come throughout them, our Advocacy staff notices them, and our prospects flag them. We imagine that making these seemingly small enhancements whereas resolving bugs can in the end have a huge effect on the consumer expertise, which is an enormous objective for our staff in the mean time. However this week wasn’t nearly bug fixes — this was about taking a chance to convey the entire engineering staff collectively to enhance how we work. We spent every week constructing higher habits round releasing modifications, enhancing our communication, and collaborating extra carefully with our Buyer Advocacy staff and one another.
How Fixathon labored
Forward of the Fixathon, we crafted an inventory of duties for our staff to work on fairly rigorously to make sure the chance of outages was as small as attainable. From there, we wrote up a set of tips for all of our engineers to leap proper in as soon as the Fixathon began.
To begin, we opened up a software referred to as Bugsnag
Every day began with our engineers participating in Bugsnag, a software designed for monitoring and managing automated error studies. The first goal throughout this section was to triage these automated error studies. Our staff reviewed, categorized, and, the place attainable, resolved points immediately reported by Bugsnag, and our objective was to carry out an motion on at the least 15 bug studies earlier than shifting to the following section. This course of was important for understanding the state of our programs and attainable issues in them – particularly since we have been responsible of not doing that on an ongoing foundation earlier.
Subsequent, we switched to product bug studies and resolved them
After the morning spent triaging Bugsnag studies, our engineers shifted their focus to the product bugs a few of them pre-selected earlier with the assistance from the product staff. These have been points beforehand recognized and logged in Jira. Engineers chosen duties within the areas they have been accustomed to and made positive to work on them safely, reviewing one another code and double-checking all the pieces was performed in a correct solution to keep away from any points on this delicate time.
We then saved the Advocacy staff within the loop
An enormous a part of the Fixathon was ensuring our Advocacy staff was conscious of the modifications we have been making. We posted each day messages into an updates channel with a roundup of the day before today’s fixes.
We spent the final day of Fixathon on high quality assurance (QA)
As a substitute of constructing extra modifications on the final day of the yr earlier than closing for the vacation season, we spent a day ensuring we have been leaving issues in an excellent state with some QA. We have been going by our product and ensuring all the pieces works as anticipated, spent a while collectively within the calls to catch up, and simply have a jolly ol’ time.
The outcomes of our first Fixathon
Finally, we realized a ton by doing a Fixathon. We realized methods to higher function in Bugsnag as a staff to scale back the numbers, we obtained higher at collaborating with one another and Advocacy, and we gained visibility into many points that in any other case would go unnoticed. We’re hoping this course of stays with us, and we are going to clear new Bugsnag points on an ongoing foundation, making it simpler for us to handle our bugs and see them earlier than our prospects can do it!
The opposite a part of Fixathon was fixing precise bugs! We resolved 44 duties complete, a lot of them small bugs we didn’t have time for earlier within the yr. Whereas most of them have been so small that it’s arduous to note the change individually, all of them contribute to our imaginative and prescient of a consumer-grade expertise.
A couple of notable bugs we squashed included:
- Including a billing web page hyperlink to the account settings dropdown
- Making “urged media” scrollable
- A number of styling modifications to make our margins and look constant all through the product
As part of the Fixathon, we’re additionally near releasing the “first remark” performance for LinkedIn, which we have already got for Instagram. (Be part of our beta program to get first entry.)
We crafted the checklist of duties to work by fairly rigorously to make sure the chance of outages was as small as attainable. And what’s most necessary – we delivered; there have been no Fixathon-related incidents previously weeks!
We think about Fixathon a fairly profitable initiative, and although it’s the primary time it has occurred, we hope it’s not the final one! We squashed bugs, spent a while collectively, and realized cool issues. And what’s most necessary – we improved our product in consequence!
