Fine, later on then, but when does the dust ever settle around here?! In the meantime I offer this suggestion. General region based host names that average miners would typically use could redirect in the case of server outages (eu, us, na-north america if including canada I suppose, whatever) More specific city, state, or country server host names stick to their respective servers (ny, sf, nl, montreal, etc.) Just post the expected behavior for each host name on the web page server list.
My primary goal here is uninterrupted service and simplicity of use. Default behaviour needs to be as simple and requiring as little user configuration as possible, so I probably won't be changing it. Having region names (eu, us, etc) switching in case of server outages but server names (ny, sf) not switching will make it confusing for users who don't read detailed instructions, FAQs or hundred-page threads (basically most of users). I'm going to stick with consistency here and switch subdomain names as necessary.
Or give us alternate sticky server location host names if that's preferable to you. It's obviously a bad practice for us to hard code ip addresses into our configurations as in light of ddos attacks they might need to be changed with little notice, but there is a comfortable middle ground somewhere to allow informed miners the ability to choose their own destiny during server outages, while simultaneously moving the rest of the herd over to a functional clevermining server.
This is something I could do and more informed users who'd like to have more control and can spend more time setting up failover plans could use it. I can provide alternate region names which won't be switched to another regions in case of outage (e.g.eu-fixed, na-fixed). I'm not sure though if there will be many people using it besides you.
That second option works just as well, and as long as there is some host name that does not move around, then that would be great.
And I would not underestimate your miners, Terk. Some of the ones who understand what is going on are likely to have greater hashpower at their disposal, and are probably tweaking for maximum efficiency, including pool server failover priorities, as they have made larger investments. And as many people are more polite than I sometimes choose to be, they might just be glad that I am asking, so they don't have to.
In my opinion, the perfect multicoin pool would only capture a miner's hashpower for as long as it believes it can be more profitable than baseline estimated earnings (in this case litecoin), and allow miners to automatically fail over elsewhere else after not finding any profitable opportunities for 5-10 minutes or so, but I understand that it is more profitable for a pool operator not to that, and how you might take a different position on this. It is your pool so of course you can do whatever you want to, but sticky locality based server host names does not sound like much to ask for, as additional unexpected latency can affect miner efficiency.
