[nemomarx]

It's about risk profile. The question isn't "which region goes down the least" but "how often will I be blamed for an outage."

If you never get blamed for a US east outage, that's better than us-east-2 if that could get you blamed 0.5% of the time when it goes down and us1 isn't down or etc

[Esophagus4]

But ise1 is down 4x more than use2 (AWS closely guards the numbers and won’t release them, but that is what I’ve seen from 3rd party analysis). Don’t you want your customers to say, “wow, half the internet was down today but XYZ service was up with no issues! I love them.”

I can’t tell if it’s you thinking this way, or if your company is setup to incentivize this. But either way, I think it’s suboptimal.

That’s not about “risk profile” of the business or making the right decision for the customer, that’s about risk profile of saving your own tail in the organizational gamesmanship sense. Which is a shame, tbh. For both the customer and for people making tech decisions.

I fully appreciate that some companies may encourage this behavior, and we all need a job so we have to work somewhere, but this type of thinking objectively leads to worse technology decisions and I hope I never have to work for a company that encourages this.

Edit: addressing blame when things go wrong… don’t you think it would be a better story to tell your boss that you did the right thing for the customer, rather than “I did this because everyone else does it, even though most of us agree it’s worse for the customer in general”. I would assume I’d get more blame for the 2nd decision than the 1st.

[throwawaysleep]

> Don’t you want your customers to say, “wow, half the internet was down today but XYZ service was up with no issues! I love them.”

See any companies getting credit for it in the last AWS outage? I didn't. My employers didn't reward vendors who stayed up during it.

[Esophagus4]

We got credit for it.

Shame about your employer, though.