Hacker News new | ask | show | jobs
by Spare_account 1739 days ago
>We were fortunate in that we had a backup payment gateway integration "just in case"

This seems like the key point here. I'm not a software guy or even a payments guy, I'm a network infrastructure engineer.

For anything that we want more than 99% uptime, we put in two of everything, sometimes more. Two separate service providers, ideally coming down different physical paths where practical.

3 comments

"Treat infrastructure as possible failure points and prepare accordingly" holds just as well for the payment infrastructure. Interesting and fresh perspective, thank you for sharing.
It's not possible to increase availability with redundancy in all cases, because not all financial actions are idempotent.

For example, sending money via a banking wire. If the bank goes down, you can't send a second wire through another bank without loss because the first wire is not retractable.

Great practice, specifically determining how much redundancy you need, and making sure it is available. The theme also goes with your acct name quite nicely!