memes

I work in the industry (not MSFT) on cloud reliability so I have insight.

• The cloud itself is not as stable as some people may think. Standard is 99.9 % uptime. Which sounds great, but if your DB is 99.9 % and your compute is 99.9% and your storage is 99.9 % and your network and so on and any one of those going out breaks your application, then you don't have 99.9 % but (99.9 %)^n which is a lot less impressive. You can make things fault-tolerant through redundancy but that comes with great complexity which can cause outages of its own.
• Apps like teams, like most B2B bullshit, provide added value by bundling together as much shit as possible. Chat, calls, calendar, spreadsheets, you name it. So now any one of those features going out individually can impact the whole app.
• Every one of those features in the bundle is managed by a different team, possibly in a different country and coming from a different company acquisition. So now you have to glue unrelated tech stacks together which is super expensive.
• The way you bundle things together in a SPA in a corporate environment with finite resources is by basically bundling together a bunch of iframes. Ever notice that the calendar tab on teams sometimes tells you to refresh your page to get new credentials? That's why, this fucking thing bundles its own authentication lib and barely talks to MS Teams so it can't properly refresh its tokens! If you like having one product's technical debt, now think about having 20 products' technical debt all conveniently forced to interact together in one web page!

Honestly I'm impressed by how well teams works with the very severe constraints they clearly have. Shit's got more moving parts than Ryanair's entire fleet and it only breaks once in a while.