Upcloud UpCloud Feedback?

London upCloud is reporting Harlow, close enough, closer than some 'london' airports

Code:

City    Harlow
Latitude    51.783298492432
Longitude    0.16670000553131

Between London and germany over 'private' network I'm seeing a 20ms avg ping and 2ms out to 8.8.8.8, linode 1.8ms.

I'm in the camp of linode at the moment is good enough, but I'm definitely hunting. Support has been spot on, fast and helpfull, almost human like . Often they have sorted the problem before I discover an issue.

I'm a lone wolf, self employed developer who likes to go climbing and soloing at that. I don't want, nor carry my phone on some days.

So for the moment, linode is still my default real production provider. I've also used digital ocean for secondary sites, those that if they are down. For everything else, it's on a paas, more so if the client does not have hosting budget, because they see web hosting from godaddy for 99p .

For the last year or so I've been looking for a cost effective hosting site that comprises ( no order ):
UK>EU owned.
Multiple DCs> One good site.
Virtual > Colo
Private Networking/Firewall appliance.
24/7 responsive support.
Optionally managed services.

Cost isn't on there, because if it's right, then it should be paying for it'self, somehow.

My logic for sticking on virtual vs trundling off to OVH et al, is, if the baremetal flutters, they ( provider ) know about it and will fix it sooner than I could on my own baremetal ( Got that t-shirt; zfs on solaris, superb, unless your provider jimmy's offspec drives in ).

Again, not sure how I stumbled upon upcloud, but the European owned, multiple DC, private network makes life a bit easier on some counts. A niche player might also alleviate ( falsely ) some network traffic challenges.

The problem before HA is the old chesnut, cost. I could do with a 20Gb ish db box, and that isn't the cheapest of things to copy around for a bit of peace of mind...

So the problem is, doubling up (HA)x2dc isn't economic, comforting but not economic for most, unless you don't want your phone to ring when you're 400' up a rock face . So I want a solid, on the ball co.

If linode pull through on the network upgrades I'll be back to 3/4 nines. It is, would be, nice to have a grand stack all there, tolerant etc, should I break a leg tomorrow. That said a single 'box' is still the easiest to deploy per php project. I don't really do php though and ftp'ing code makes me cry. for centminmod then

So my risk profile has softened, and for 'virtual' hosts, more so. I like to establish the impact of noisy neighbours or physical host issues on potential downtimes. Not done anywhere like enough testing on upCloud to form a view yet. Although resizing appears to migrate hosts almost seamlessly, supporting some confidence if a physical host had a problem, we'd not really see much of it up top. So my gut is telling me, keep going, you might like it!

The only sticking point and as eluded to above is ram. $ for $, ram is better on linode. For php stuff, I don't need much ram. But a ruby web app can easily gobble up 500+ on boot for a single threaded process and leak like a sieve..so more is always nice.

If I can make upCloud work, and after a month my test instance hasn't gone walkabout, I'll be happier.

I'll be happier when we can deploy our front end code to global cdn's and only have to worry about our api service Don't think XF or IPB are there yet! Not quite signed up to serverless architectures, lambdas and all that... feels like I'm writing my own epitaph if I encourage that malarky.