This gives me bad flashbacks to Salesforce's GovCloud implementation (likely mandated here for Slack), which is a steaming nightmare mess of never-quite-consist

This gives me bad flashbacks to Salesforce's GovCloud implementation (likely mandated here for Slack), which is a steaming nightmare mess of never-quite-consistently-working disjointed CI servers (multiple per project), manual deployment processes requiring 48-hour lead times for submitting requests for approvals to deploy (with boilerplate paperwork manually filled out about the artifact being deployed and the timeline of the staggered deploy), and mandatory multi-day staggered deployments with no automatic rollback, where all deployment steps were required to be performed (again, manually) outside of US business hours (no deploys between 8am eastern and 6pm eastern will be approved, even for emergency fixes). I fled from overworked stress-panic when that nightmare was forced on me because Salesforce decided to double my acquisition's workload by reorg'ing out 70% of our engineering staff (while requiring the remaining 30% to pick up and maintain — and in some cases, bring into compliance — the ownership areas previously belonging to that other 70%). Not a day has gone by that I've regretted it. Every night that I can sleep well instead of being kept awake by Salesforce-induced stress is a night I'm glad to have back from them.