Performance Degradation with App42

Question

Hi, We have noticed a significant degradation in performance over the last 36 to 48 hours with all api calls to Shephertz. Queries that have been running in 1 to 2 seconds are now taking 20 to 30 seconds and even longer. What is strange is that the billing and metering section does not show any high response times. Most queries are 100ms or lower which would leave one to believe that it is either network related or our queries are being queued on your servers for some reason. I have tested via wi-fi, wired ethernet and cellular on three different networks and the results are the same; all network activity outside of Shephertz is working normally but api calls to Shephertz are returning very slow. Any help is greatly appreciated! Thanks, Scott

Comments

UPDATE: At the startup of my app, I always do a quick check to see if I have network connectivity to "api.shephertz.com" on port 443. I started logging this connection this morning and can see that when the issue occurs it is taking 5+ seconds to make this initial tcp socket connection. When things are moving along normally, this connection is sub-second (.2 or less). This now seems more like a network issue than anything. I continue to run connection tests on different devices connected to different networks and I am only seeing these issues when trying to connect to shephertz.com. For example, this same socket connection is running sub-second every time when connecting to google.com, microsoft.com, apple.com and yahoo.com. As a note, I will have periods of 5 or 6 minutes where everything works normally then it goes back to being slow for 20 or 30 minutes. Hope this helps!

---

UPDATE2: It is around 24 hours later (Sunday) and the issue is still occurring. Below are the results of an online tcp/web connection test from various locations across the world over to shephertz.com. As you can see it is showing slow connection times. As before, there are small amounts of time (usually 5 minutes or less) where things move along quickly as it always has in the past but then it goes back to being slow.
 
Beijing, China 10/09/2016 04:00:08 PM 22092ms
Sydney, AU 10/09/2016 04:00:08 PM 9735ms
Copenhagen, Denmark 10/09/2016 04:00:07 PM 10760ms
WA, USA    10/09/2016 04:00:07 PM    6290ms
CA, USA (IPV6) 10/09/2016 04:00:07 PM 110 Failed DNS Error 12007
Mumbai, India 10/09/2016 04:00:08 PM 9328ms
Warsaw, Poland 10/09/2016 04:00:07 PM 10268ms
Paris, France 10/09/2016 04:00:07 PM 9554ms
South Africa 10/09/2016 04:00:08 PM 11459ms
Buenos Aires, Argentina 10/09/2016 04:00:16 PM 31767ms
Shanghai, China 10/09/2016 04:00:08 PM 10013ms
Tokyo, Japan 10/09/2016 04:00:08 PM 8318ns
Amazon-US-East 10/09/2016 04:00:07 PM 6398ms
VA, USA 10/09/2016 04:00:07 PM 6921ms
Tel-Aviv, Israel 10/09/2016 04:00:08 PM 11395ms
Amsterdam, Netherlands 10/09/2016 04:00:07 PM 3225ms
TX, USA 10/09/2016 04:00:07 PM 4762ms
Brisbane, AU 10/09/2016 04:00:07 PM 6174ms
CO, USA 10/09/2016 04:00:07 PM 8160ms
Frankfurt, Germany 10/09/2016 04:00:09 PM 9505ms

Answer 1

Hi Scott,

Apologies for late reply.

I have replied to your query over support. Please share your available timings and Skype details over support so that we can connect tomorrow(12 Oct) for the quick resolution of this query.

Regards,

Rajeev Ranjan