Maybe but it’s still disappointing… I recall a Monzo post early in the Mondo beta talking about how the whole system was being designed and built from the bottom up to ensure stability and avoid the outages of legacy banks. That doesn’t seem to have worked.
We haven’t seen the blog post yet, so we have no idea what failed and if it was Monzo tech or not. Best not to jump to conclusions without facts.
As one of the people that does a lot of work on our social media, I wouldn’t mind investigating this for QA purposes.
If the tweet is still in your recent replies, would you mind DMing me a link?
Some of the tweets from relevant team members inside Monzo indicate that they were not exactly delighted that the (partial) outage occurred
However the infrastructure design did work well in some senses - they rapidly knew there was a problem (leaving aside concerns around Twitter for a moment), they were able to roll it back quickly and got things running. Not ideal that it took a couple of hours, but this was not the multi-day dramas of RBS in 2012
Looking forward to finding out exactly what occurred. Been boning up again on the Monzo infrastructure in the meantime to maximise my understanding of the writeup
We’ll have more to share soon but this particular outage was entirely with us.
Potentially harsh, putting aside Faster Payments (which go through a 3rd party), card processing has generally been extremely stable since we brought it in-house and on our own systems.
Even with this most recent incident, card payments did not experience a 100% outage like were frequent in the prepaid days (I know that’s no consolation to those who had transactions that were affected).
Systems such as ours are not always designed to prevent failure, they’re designed to quickly recover from failure. Particularly, because we own the entire stack, we have engineers who know the services they own inside out and can quickly ship fixes where problems do occur.
Looking forward to it
Thanks for the (I think first) confirmation that it was solely an internal issue. Not ideal obviously, but good to be able to admit it, “own” it and go from there