It seems like a bunch of the big tech companies are shuffling compute around rather than getting more. Are we hitting the limit of the currently available compute, or is this just MS being bad at things?
post
Look, GitHub is an entry-level offering for the general public. If you want true worst-in-class performance you have to go with Azure DevOps.
GitHub simply doesn't have features like build steps that randomly take ten times as long as usual to complete for no reason, an inability to refer to a PR just by its ID even though it's unique, or a navigation that makes no sense at all. Sure, you get that four-eights availability but do you get pipelines that reliably fail even though they worked just fine yesterday and nobody changed anything? Those are enterprise features for companies that pay big money to get the worst damn Gitlab clone money can buy.
If Azure had an 11% downtime, then I'd consider it a blessing. 19 hours a week when you'd be free of using the damned thing. But no, it just stays up and taunts you.
But does it really stay up, or does it just suffer random undocumented service degredation near constantly? ;)
Can't tell you how many times I've built some basic automation off the fact that certain operations through Graph are effective basically instantly, only to have to add increasingly long timers between steps like a week later when Azure suddenly decides that it's done being "nice".
all 4 comments