
We spent the last hour in a meeting about setting up SLOs. The SLO guys were initially talking about tradeoffs between new features and reliability, until we pointed out that this thing is going into maintenance mode at the end of the month and is definitely not getting any new features before then. So it all just devolved into setting up some monitoring so somebody gets alerted if it starts running really slowly

It did make me realise, though, that we haven’t got anything set up to notify us if it just dies. Up to now, we've relied on the fact that the users would send us a message on Teams saying “Why is it dead?”

But strictly speaking, it’s other people’s responsibility to notice that level of infrastructure failure and sort it out. They get paid to be hassled by PagerDuty; I don't

It did remind me of the fact that the BBC iPlayer people have, or had a few years ago, a thing that alerted them if iPlayer mentions on Twitter suddenly increased, as that usually meant it had fallen over






.
.
.

Leave a comment: