Google Meet Went Down and Everyone Panicked
Photo by Rubaitul Azad on Unsplash
Incidents are just part of our work. The product doesn’t work? Have a meeting about it. A user in Albania is being given US terms and conditions? Raise an incident.
So when I7823 came up, many people had a quick look at the Slack thread.
Google were having an outage. Perhaps not surprisingly because Cloudflare were having an outage. The result?
A meeting, an incident meeting of the best and greatest that we have to offer.
“When is it going to be fixed?”
“Nobody knows, it’s CloudFlare”
“What is the mitigation?”
“None, CloudFlare is down along with most of the web services we use”
“So when can we fix it”
Brilliant. When in doubt keep asking everyone when it will be fixed and we’ll get a solution. That and what we should write in the “communications” to our management team.
Oh. And
When will it be done
When will it be done
When — will — it — be — done?
The Illusion Collapses
I often think that my elders and betters are indeed my superiors. I’m so wrong that it hurts, they don’t really move things forwards or improve things for us on the ground floor codebase.
So when Google Meet goes down they panic. They don’t know what to do and wait for “engineering” to come in and pick up all of the pieces for them. It seldom happens.
I find that my illusion is shattered. Instead of “best-in-class” I find that we have a group of semi-idiots who are unable to fix the moving parts that don’t so elegantly fit together.
The first sign of trouble is suddenly panic. How are we going to have our catchups? Won’t this incident spawn a set of new incidents?
You can almost hear project managers crying in corners of the office. Which is kind of surprising, since almost everybody is working from home on this particular Thursday (it’s up to you to discover when you think this event happened).
The Result
The meetings stopped.
Everything carried on (apart from those poor souls caught in the incident meeting themselves).
We didn’t have a retro, so that didn’t help (?!). People just… worked.
Pull requests appeared. Bugs got fixed. Documentation was updated. One developer I hadn’t heard speak for six months suddenly merged three tickets before lunch.
It was like discovering an ancient civilization beneath the office floorboards.
I’ve written before about how software companies confuse communication with productivity, and this was a demonstration of how many developers now spend half their working day explaining what they are about to do instead of actually doing it.
The outage accidentally removed the ceremony.
It turned out the ceremony was the bottleneck.
Nobody was able to lean on meetings for reassurance.
Managers couldn’t “see” people working.
It was a version of heaven
Conclusion
Children at school get more autonomy than some senior software developers.
I’ve worked in teams with multiple standups a day where nobody listened to anybody else anyway.
Half the people are on mute.
A quarter are making coffee.
One person is clearly shopping on another monitor.
Someone inevitably says “sorry, what was the question?” after twenty minutes.
Yet companies continue adding more meetings as if communication debt can be solved through brute force.
I experienced what it was like to get rid of this all. It was really good, as a matter of fact.
Of course, by the next day everything returned to normal. Very little got done.
About The Author
Professional Software Developer “The Secret Developer” can be found on Twitter @TheSDeveloper and regularly publishes articles through Medium.com
The Secret Developer once attended so many meetings in a single day that they became nostalgic for compiler errors.