> the whole point of using a cloud provider is to lower the amount of ops work
In a sense. You shift a lot of what ops work used to look like to someone else. No longer you have to worry that there is some issue with a fiber optic transceiver. This is someone else's problem.
This is an abstraction though, and abstractions leak. Even with the best teams and monitoring available, they are not going to catch it all. But at least you can see if there is an issue, instead of filing a ticket to even find out if there was an issue to begin with. And they won't able to tell you that "everything looks fine in our end".
At the end of the day, it is still their problem to fix.
I think this is a pretty bold move, I wouldn't want to expose internal data like this to end customers, unless there was complete confidence on the monitoring systems. Issues with data have the potential to generate an immense amount of grief from customers.
In a sense. You shift a lot of what ops work used to look like to someone else. No longer you have to worry that there is some issue with a fiber optic transceiver. This is someone else's problem.
This is an abstraction though, and abstractions leak. Even with the best teams and monitoring available, they are not going to catch it all. But at least you can see if there is an issue, instead of filing a ticket to even find out if there was an issue to begin with. And they won't able to tell you that "everything looks fine in our end".
At the end of the day, it is still their problem to fix.
I think this is a pretty bold move, I wouldn't want to expose internal data like this to end customers, unless there was complete confidence on the monitoring systems. Issues with data have the potential to generate an immense amount of grief from customers.