needaserver.net route problem, this normal?

notgod

Hi
3 servers in needaservers.net
i use HK/JP/USA servers reverse proxy to needaserver.net servers
recently it's very slow in last week, but before previous week not problem.

so i test it use MTR in much location
3 x HK
1 x USA CA
1 x Canada
1 x Japan tokyo

**the result **
if images small click this Imgur Imgur

**here is explanation from NEEDASERVERS.NET **

Hello
We have no issue, this is simply de-prioritization of pings, this is done with our routing gear to maximize the use of the CPU of the gear- We de-prioritize ICMP tp conserve the memory of the routers.
The packet are being passed just fine- the impression of latency, is not latency is just that our routers and this is very common to cast of this traffic in datacenter- packets are being passed along fine.
Victor Goodman

so i want asking this packet loss is normal?
the core packet loss 50% it's normal?

PrincessOfCats

@notgod said:
Hi
3 servers in needaservers.net
i use HK/JP/USA servers reverse proxy to needaserver.net servers
recently it's very slow in last week, but before previous week not problem.

so i test it use MTR in much location
3 x HK
1 x USA CA
1 x Canada
1 x Japan tokyo

**the result **
if images small click this Imgur Imgur

**here is explanation from NEEDASERVERS.NET **

Hello
We have no issue, this is simply de-prioritization of pings, this is done with our routing gear to maximize the use of the CPU of the gear- We de-prioritize ICMP tp conserve the memory of the routers.
The packet are being passed just fine- the impression of latency, is not latency is just that our routers and this is very common to cast of this traffic in datacenter- packets are being passed along fine.
Victor Goodman

so i want asking this packet loss is normal?
the core packet loss 50% it's normal?

Packet loss is normal, ICMP is deprioritized on many routers. As long as the endpoint has no packet loss, this is fine.

The pings do look a bit high though. - read wrong mtr, ignore that

notgod

the problem the web site loaded at least 3 times slower

PrincessOfCats

@notgod said:
the problem the web site loaded at least 3 times slower

From where - all of the nodes?

notgod

@ALinuxNinja said:

@notgod said:
the problem the web site loaded at least 3 times slower

From where - all of the nodes?

yes, 100% sure not change node and backed servers,
use webpagetest speed and check nginx logs , all is waiting in hk/jp => needaserver backed
i use hk/jp servers local try curl test same result

but i move some data to other LAX DC( like psych.net sharktech)
not this problem , very fast.

so 99.9% sure the problem in needaserver network

and my another one needaservers windows servers, the network very unstable
some times windows RDP always disconnect, can't ping IP, network offline
need wait a few minutes restore.

encountered many times, some customer say ARP problem,NO VLAN

notgod

@ALinuxNinja said:

@notgod said:
the problem the web site loaded at least 3 times slower

From where - all of the nodes?

and start in last month, my monitor pre day tell me down down down down much times
max day down over 50 times , pre time few minutes

like this