Not Santander as well! They were usually fine when I was with them. Just another example of this becoming more and more common.
I don’t think it’s becoming more common, it’s just more people are using online & mobile banking, and more often, so it is more noticable.
Every bank (even the new ones) will have downtime.
In areas out of their control yes but their apps no. I don’t know about monzo but the starling app has never been unavailable in any way.
Just because they haven’t been down doesn’t mean they won’t ever go down.
From what I know you might be waiting a while
I shouldn’t be inevitable. With good design and redundancy built in
I’m struggling to think of an IT system that has never been down.
Can something really be down, if it’s never truly been up?
Is that Starlings secret?
I ran stuff for quite a while and never had any downtime. Just freeze the attached dependencies, make sure your system isn’t easy to.
Then do the following:
- Use some form of supervisor that can tell if the program has been killed and kill a program when it uses x amount of memory
- have the supervisor restart it after the program dies
I’m not talking about individual servers here, I’m talking about IT services of the sort needed to run a bank.
I’m well aware of servers, and indeed have worked on some, that have had many years of OS uptime - but that doesn’t mean that the services that rely on those servers share those impressive uptime statistics for a multitude of reasons.
And that supervisor method is just masking issues with an unrealible application.
Of course, you are right, the best way to avoid downtime is to not change anything, and if you’re (very) lucky you’ll have no hardware or network issues and your application has no bugs.
But, in reality, you will need to change stuff (particularly if you’re a fintech bank!), and you’ll have all of these issues happen at some point and that’s when keeping things up gets tricky.