Heh, that's why we open sourced our platfrom+devops. Check it out - its totally next-level magical stuff https://github.com/allegroai/trains

What does that entail? Are there interesting stack choices? Best practices? Pain points? Dont-use-X-in-your-stack-you-will-regret-its? Gotchas (that they'd be willing to share)?

Your platform teams monitor data quality for an upstream ETL?

All of the platform teams I’ve worked with are concerned with monitoring things much lower ok the stack than data quality.