This is exactly what I was waiting to see, great front-end, linked to a well thought-out and well-documented back end. Great work guys
To elaborate a little more on what I think makes the front end "great": I think it's very simple. I contemplated making something like this simply because I think tools like nagios,etc are swiss army knives, BUT they look really ugly. I don't see why we can't have some beauty in monitoring.
I think that beauty, and function actually come together well in the interface as it stands now (I know it's not complete) -- I immediately understood where to look, what to click. While the everyday-user is not going to be unfamiliar with the interface, I also think an experienced user could benefit from having the clarity of a detail view, and individual machine view (the right and left respectively).
One place I can think of that might not be ideal for the kind of design is if you want to know aggregates/historical data of a certain data center/machine -- but I think this could easily be remedied by adding a stats page that is made to show that kind of broad data specifically (or even keep it out of the core codebase, and just have it be a very very popular plugin -- small tools ftw)
After last seeing Consul posted 2 weeks ago, and playing with Docker for deploying my application, it got some gears turning. I think the two would be a great combo. Though I like the philosophy of immutable containers that Docker asserts, where if the topography needs to be reogranized (with linking containers [1]), you can just tear them down and spin them back up in a second. In practice this is kind of a pain, and a very top down view. Using Consul would make it very nice for snapping together infrastructure, with more robustness provided, i.e. health checks & discovery.
As a side note, I played around with Ubuntu's Juju a bit, and they also allow you to change an instance's configuration during runtime. But it's from a top down perspective, than distributed.
I just found out about CoreOS a couple days ago. I don't think it's necessarily what you're talking about, but seems a little similar https://coreos.com/
Do you mean service discovery as in UDDI? It sounds a bit like it's old SOA ideas all over again - they failed and it would be beneficial to understand why before we make the same mistakes.
Not sure, I've not heard of UDDI before. Mostly I mean, in a really general sense, that servers come up and become part of a cluster or group of systems providing a service.
says Microsoft, IBM, et.al. abandoned their efforts of providing public service discovery infrastructure. Doing it the same for REST may just be repeating of old mistakes.
UI isn't responsive to my not really that small laptop screen. I have to horizontally scroll to see everything when there is little to see. There is too much whitespace everywhere.
To elaborate a little more on what I think makes the front end "great": I think it's very simple. I contemplated making something like this simply because I think tools like nagios,etc are swiss army knives, BUT they look really ugly. I don't see why we can't have some beauty in monitoring.
I think that beauty, and function actually come together well in the interface as it stands now (I know it's not complete) -- I immediately understood where to look, what to click. While the everyday-user is not going to be unfamiliar with the interface, I also think an experienced user could benefit from having the clarity of a detail view, and individual machine view (the right and left respectively).
One place I can think of that might not be ideal for the kind of design is if you want to know aggregates/historical data of a certain data center/machine -- but I think this could easily be remedied by adding a stats page that is made to show that kind of broad data specifically (or even keep it out of the core codebase, and just have it be a very very popular plugin -- small tools ftw)