<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">That said, some projects/tools should be moved off into their own labs<br>
projects (quarry and wdq come to mind), but trying to build an<br>
alternative tool labs isn't going to work.</blockquote><div><br></div><div>I by no means would like to do such a thing. I was just suggesting it as something that seems convenient as a consumer of labs resources. If it makes sense for ops to monitor one cluster then that's ok by me. It just seems like sometimes we have to do things that we're not 100% sure are safe, and when they're not, some of our tools go down. Since a lot of people depend on those tools having pretty much 24/7 uptime, and since those same people usually seem to not mind if they're not running on the fastest labs hardware, best NFS setup, etc, I made the "graduate labs" suggestion. But this is because I'm 100% ignorant of how ops works :) Feel free to interpret my point of view however you wish, but I am not trying to cause ops more work, just stating things as I see them. </div></div></div></div>