Akamai Offload reduction since the New Year?
This isn't an outage per say, however it is a pretty major reduction of service availability that we've seen, and perhaps other folks on this list have seen something similar. Since literally 00:00:00 2012/01/01 our Origin Bandwidth Offload has been slowly dropping. Prior to 01/01, we were seeing 97% offload from our origin servers (not bad!) Midnight 01/01/2012 Origin offload dropped to 95% and has been slowly, steadily dropping since. Today we've reached a new low of 88%. We've contacted Akamai and so far we've received very little feedback about the cause, or why this happened so dramatically at midnight on January 1st. Our origin infrastructure has not changed before or after 01/01. THe lack of willingness to investigate on Akamai's end has caused me to reach out to this list. Has anyone else experienced something similar? If so would you please contact me directly? I would love to hear your situation. Thank you very much, - S
On Feb 17, 2012, at 7:06 AM, dl wrote:
Has anyone else experienced something similar?
I've seen something similar whereby HTTP attack traffic which has been deliberately constructed to be non-cacheable has been relayed from a CDN to the targeted site. Have you looked into your HTTP query details in order to see if you're getting lots of queries which stand out in that regard vs. your regular traffic? Another possible cause would be changes to a site itself which results in more non-cacheable dynamic content which can't be served via the CDN. ----------------------------------------------------------------------- Roland Dobbins <rdobbins@arbor.net> // <http://www.arbornetworks.com> Luck is the residue of opportunity and design. -- John Milton
So far we have seen, all requests have been legit. Our content is very static binary files, which makes this a little more confusing. It almost feels like our service has had its expiry adjusted on the Akamai side, causing our content to expire faster than it use to from edge servers. Or perhaps a new customer has come in and is taking over the majority of the cache on these same servers? Impossible to tell, and not information Akamai wants to share with other people. Thank you all for your contacts. I'll be happy to hear anything else anyone wants to suggest. On Thu, Feb 16, 2012 at 16:27, Dobbins, Roland <rdobbins@arbor.net> wrote:
On Feb 17, 2012, at 7:06 AM, dl wrote:
Has anyone else experienced something similar?
I've seen something similar whereby HTTP attack traffic which has been deliberately constructed to be non-cacheable has been relayed from a CDN to the targeted site.
Have you looked into your HTTP query details in order to see if you're getting lots of queries which stand out in that regard vs. your regular traffic?
Another possible cause would be changes to a site itself which results in more non-cacheable dynamic content which can't be served via the CDN.
----------------------------------------------------------------------- Roland Dobbins <rdobbins@arbor.net> // <http://www.arbornetworks.com>
Luck is the residue of opportunity and design.
-- John Milton
_______________________________________________ Outages mailing list Outages@outages.org https://puck.nether.net/mailman/listinfo/outages
participants (2)
-
dl -
Dobbins, Roland