March 31, Caravan. UPD

A couple of days ago, the Caravan sent out a warning to all customers about planned work on March 31 at 18-30. At 16, with a penny, as reported by tech. Support, an accident occurred with power.
Projects such as Habr, Lepra, Free-Lance and so forth lie and do not move half a day ...
Time 20-57, Habr came to life, part of our servers came to life, part continues to lie. Including DNS. So hello 2 o'clock in the morning and not earlier.
Several cloud services are running on our servers. The loss in 4 hours is about 10k rubles. Not to mention the moral damage and the blow to the reputation.
At the Caravan this is not the first time, and to be honest, it is annoying! Ponty and prices like a really cool telecommunications company. However, whoever was in the DC and saw this collapsed trash, sovkovy plant, obsannye stairs, etc., he understands that this is just a show-off.
From the alternatives we look in the direction of Oversana. And the prices are much nicer (apparently stock) and DC modern. Anyone else have any ideas? It would be interesting to hear the opinion of the owners of the habr, as they relate to such. Somehow it is not normal that ...

ps But I was more shocked when we rented a cisco guard service from them for half a year, and once fell from insignificant DDoS. Specialists of the Caravan spread their hands and said that the tsiska was not configured. On my request to set up, they twirled something there, the mobile version of our service (which is an integral part of the service) stopped working. It turned out cisco decided that this was also left traffic.

UPD. 23:48, I am in METRO, I buy potatoes and sausage. The partner is calling, telling me to seize the server power supply at the same time in the metro, as our on the go burned out. It turned out before bedtime he called the caravan again. The dialogue was about the following:
- Greetings. Again I, the contract is such and such, the company is such and such. Some servers are still not pinged. Interested in such and such.
- Just a minute ... Oh, and you have serious problems, along the way the server burned out, the lights are off.
- No, guys, you have problems ... Write out the passes.
')
Broke, go. Mood was otpi quarrel with this "tech support." It struck not only amebic, but the complete lack of desire to enter the position or at least somehow help. Previously, we decided that the server was alive, that the BP had burned down. At our requests to provide BP on the spot, we were refused. Buy is the same. I remind you that we have cloud services around the clock, 150 online companies are out of work, they blew up the phones of our managers. It is on one side. On the other side was technical support for the caravan, which did not want to help us find the BP at night. Not that it did not help, but stupidly did not want to help.

We arrived customer clogged. While we were sorting out with our server, person 2 changed at the adjacent tables. But the problem is the same for all - the servers just did not come out of the reboot. The lights are on, everything is ok. Someone bios cursed at frequent failures, it was necessary to press F2. We frya refused to work without fsck in manual mode. And all would be nothing if it were not for "Oh, and you have serious problems." Although in general already in the client we talked with the “colleagues”, we were informed that we would wait about 4 hours for KVM. But we would prefer to wait for KVM, if not "Oh, and you have serious problems."
All those who were with us in the client quietly precipitated when it turned out that all was well. Everything was all good, but everyone came to the call, "it seems your server burned out, the lights did not blink."