Some of it is that they don't want to alarm customers unnecessarily, and wait until they have more information about what is going on and what is being done to

Some of it is that they don't want to alarm customers unnecessarily, and wait until they have more information about what is going on and what is being done to fix it. There's no sense crying "wolf" if you do not yet know for sure that it actually is a wolf. And putting up a notice that just says "there is something going on but we have no idea what it is" will alarm customers unnecessarily and simply will not help the situation. Making the call to put something on the system health dashboard is a big issue, and needs to be carefully considered. And even if the system health dashboard might not show anything, there is likely a personal health dashboard that may show you details of something that is going on that will affect you personally, even if that issue does not (yet) rise to the level of being shown on the system health dashboard. So you should always look to the PHD first. That said, I think operators should err on the side of being more open with information. Disclaimer: I work at AWS and sometimes I get to see some of the internal discussions regarding some outage issues, and frankly I think we do a pretty good job in this space. Well, at least most of the time. But we are human, and sometimes things happen that, in hindsight, should have been handled differently. And we have a process for actively working to take those lessons onboard and try to make sure we don't make the same mistakes again.