[xmprt]

I understand this isn't a perfect world and problems happen but we also shouldn't live in constant fear that a deployment will break production. There should be processes in place that prevent broken code from even getting to the point where it's merged into master.

The biggest extreme that I've faced was a team where there was a 36 hour window between Wednesday and Thursday where deployments to production were allowed and it was a nightmare getting any code out especially with most deployments involving 10-20 new commits so if something did break we had to rollback a lot and rack our heads to figure out what went wrong.