Does anyone know why IMP could be behaving strangely ?
My imp keeps going in this weird mode where server log displays the device has been disconnected and IMP doesnt respond to anything my hard reset.
Sounds like you have an infinite loop in your imp code, causing it to not process any packets from the server and hence go offline?
I had the same problem just this afternoon. It was loading and running cod just fine and without changing anything, mine started doing this too.
How do you clear its brain and get it to work again?
Im getting “blinkup failed” LED codes too. I havent changed my code - there was some kind of server outage today which knocked my imps offline - I couldnt even get to ide.electricimp.com from two different ISPs - now the IMP is happily connected to the WIFI, but it cannot connect to the IMP server.
I’m seeing the same thing. No changes made on my end; several devices just stopped working giving orange/amber startup LEDs.
Of note, these not working are attempting to connect to production01.boxen.electricimp.com (54.241.245.144), while the ones that work are connecting to imp.electricimp.com (184.169.136.13). The connection to the former address does succeed, then nothing happens. I have a packet dump available (wireshark/ethereal/pcap format) upon request.
Connections to imp.electricimp.com are now failing (server responds with RST), so my devices still trying to talk to that are now not working, as well.
I also got bumped off of electricimp.com’s IDE but I didn’t think that it was their server. If they had a major problem, that could explain things. In any case, I can’t blinkup the board any more either, so I’ll just give up for tonight and see what things look like in the morning.
I’m glad to hear it’s not just me. This is a one-of-a-kind prototype that I’ve been working on for weeks and if it had died or some bogius code is stuck in it that I can’t recover from, I’m going to be miffed.
Jup here 2, Pushover messages woke me, i blamed my wifi, but now iam not so shure
Just pasting my log here for if they want to know the time
2014-05-27 02:33:38 UTC+2 [Status] Device disconnected
2014-05-27 02:33:38 UTC+2 [Agent] device disconnected
2014-05-27 02:33:46 UTC+2 [Agent] Pushover response: {“status”:1,“request”:“e1e74d203802d88a846****”}
ok its up again here, good night 
See @impstatus, amazon had a major outage; this was causing reachability issues (internal routing within the datacenter), disk outages, server outages and more fun. It’s still ongoing now, but Amazon aren’t admitting it on their status page - their support staff are, though.
Apologies for the inconvenience. Right at this second, we have 80%+ of production servers quite happy (and imps will fail over from non-happy servers so there’s no visible impact), but the development server that most of you are on doesn’t have a failover and is suffering from an unannounced background disk migration… and as such is crawling.