[outages] Email to Microsoft Office 365 users

[Jeremy Chadwick]

I was seeing what looked like "general slowness" when it came to
getting A, MX, or ANY back for staples-com.mail.protection.outlook.com.
Much of the delay (using +trace on dig) showed high latency in a couple
spots, but I never saw any actual timeouts:

;; Received 763 bytes from 192.5.5.241#53(f.root-servers.net) in 885 ms
;; Received 871 bytes from 192.48.79.30#53(j.gtld-servers.net) in 368 ms
;; Received 203 bytes from 65.55.37.62#53(ns1.msft.net) in 204 ms
;; Received 141 bytes from 134.170.140.11#53(ns1-gtm.glbdns.o365filtering.com) in 168 ms

Total response time was roughly 4-5 full seconds.  This was around 15:07:40
Pacific Time or so, give or take 3-4 minutes.

Presently I'm not able to reproduce the behaviour; response times seem
to be decent now (1 second tops), with the majority of the remaining
delays being here:

;; Received 203 bytes from 208.84.2.53#53(ns2.msft.net) in 305 ms
;; Received 141 bytes from 157.56.116.52#53(ns1-gtm.glbdns.o365filtering.com) in 409 ms

Nothing looks bad via mtr, except in the case of packets from me to
157.56.116.52, which appears to be going to Europe (I'm under the
impression "ams" refers to Amsterdam):

15. xe-10-2-2-0.bl2-96c-1b.ntwk.msn.net (204.152.140.87)             0.0%    47    46  83.5  83.5  81.3  95.1
16. ae11-0.nyc-96cbe-1b.ntwk.msn.net (207.46.45.85)                 37.8%    46    28  84.8  85.2  82.3  97.8
17. xe-0-3-1-0.lts-96cbe-1b.ntwk.msn.net (207.46.43.47)              0.0%    46    46 156.7 157.1 154.5 165.7
18. xe-2-2-2-0.ams-96c-1a.ntwk.msn.net (207.46.44.59)                4.3%    46    44 167.1 164.8 163.4 172.0
19. ???

The latencies here are likely normal, packet loss may be ICMP
prioritisation-related (can't tell because destination or preceding
routers don't respond to ICMP), but it does bring into question if
Microsoft has a form of geo-located GSLB (if not, ignore.  If so, why
are queries from the US west coast going to Europe).

-- 
| Jeremy Chadwick                                   jdc at koitsu.org |
| UNIX Systems Administrator                http://jdc.koitsu.org/ |
| Making life hard for others since 1977.             PGP 4BD6C0CB |

On Tue, Oct 21, 2014 at 05:01:22PM -0500, Frank Bulk via Outages wrote:
> Our last queue check reported DNS failures looking up <domain
> name>.mail.protection.outlook.com for 11 such domains.
> 
> Just picking one, staples.com: 
> 
> root at nagios:/tmp# dig MX staples.com +short
> 10 staples-com.mail.protection.outlook.com.
> root at nagios:/tmp# dig  staples-com.mail.protection.outlook.com
> 
> ; <<>> DiG 9.7.3 <<>> staples-com.mail.protection.outlook.com
> ;; global options: +cmd
> ;; connection timed out; no servers could be reached
> root at nagios:/tmp#
> root at nagios:/tmp# host MX staples.com
> ;; connection timed out; no servers could be reached
> root at nagios:/tmp#
> 
> 
> Anyone else seeing this?  A colleague says that O365 is having issues and
> will be forwarding me some details.
> 
> Regards,
> 
> Frank Bulk
> 
> _______________________________________________
> Outages mailing list
> Outages at outages.org
> https://puck.nether.net/mailman/listinfo/outages