Honored to be joined by Adam Clark and Peter Vanhee of hugely successful British charity, Comic Relief. You can read more about Red Nose Day here but, for the purpose of this technology blog, here’s the short version – they raise over £60M ($US80M+) in one evening, processing as much as 300 donations per second. No they’re not wearing capes, but maybe they should be…
I’m sure there are folk out there thinking “hey, my infrastructure can handle that”. I get it, but can you cost effectively acquire, stand up, maintain without having a heart attach, and then tear down that infrastructure without upsetting a single user/customer? Well, these guys can.
Things learned in this episode:
- DPS – understanding this hot new metric Donations Per Second
- Planning architecture for 1 evening’s use
- The level of observability it takes to ramp up to a 1000x workload in one evening with confidence. Hint: IOpipe was key
- The cultural aspect of building and running their donation service. Hint: how they communicate with each other is a big part.
Enough words form me, watch now:
Thanks for listening!
Episode #50. Wooohooo! And for this landmark 50th episode we’re delighted to have super-popular Serverless observability company, IOpipe, on the show. In this episode (video below) I’m joined by IOpipe founder and CTO, Erica Windisch.
“Unless you set your concurrency for the lambda to ‘1’, right, which you could do [LAUGHS], you’re opting into building a distributed system.”
Erica dropped out of school and joined a web hosting company in 2001. At the time she was looking for better orchestration and observability solutions but nothing came close to her requirements… so, she started building the tools she needed.
“Through all of this we were spending all this time focussing on building the best orchestration tool, and management tool, and customer experience from deployment automation but so little effort was being spent on really amazing tools to operate and manage these workloads.”
Always at the cutting edge – one of the early developers on OpenStack, and an early employee at Docker – Erica started looking at Serverless-type technologies back in 2010. On how thing differ between traditional architecture and Serverless she shares, “you get this amazing scale but you’re also opting in to building a distributed system“. This brings new complexities that IOpipe can really help with.
Watch this episode to understand the level of confidence her company gets from their offering and how “IOpipe production infrastructure monitors itself with IOpipe production infrastructure“.
Thanks for listening!