HTTP Request Failure (Error 200)

Question

I am using ArcGIS Enterprise 10.6.1. I have a Federated Server that has ArcGIS Image Server installed on it which has its own ArcGIS Server Manager.

I am using Geocortex Analytics 1.7 with the agent installed on this same server.

3 times over past 4 months at random times, I have received an alert from Geocortex Analytics that notifies:

"HTTP Request Failure (Error 200); The response from server could not be parsed; Could not access any server machine".

What happens is all the services on this instance of ArcGIS Server go down (it's like they are all restarting at the same time), and then 15-20min later the error goes away all is back to normal.

I checked the ArcGIS Server Logs, nothing related to this issue.

I checked the Event Logs on the server machine, nothing related to this issue.

I searched the ESRI Community and Help, nothing related to this issue.

There is plenty of memory and RAM available on the server machine.

As a long shot I decided to post on the community, maybe one of you have experienced the same thing or have an idea on why this is happening.

I am planning to upgrade to 10.9.1 (along with installing VertiGIS Analytics) within next 12 months as 10.6.1 is almost at end of life.

berniejconnors · Answer

I agree - we see a similar ~15 minute outage when we reboot a server.

Scott_Tansley · Answer

Have you applied windows patches or performed other maintenance on the server in that period?  The errors align with a restart of the windows services on the machine.  A restart will typically take up to 15 minutes to restart all web services (depending upon how many there are).  I typically see errors like this when the IT department applies window patches or makes changes relating to Anti-virus etc.

TimHayes3 · Answer

Thank you, Bernie. I checked the Windows Logs and found nothing that corresponds with what the issue is and the timestamps show nothing unusual at the time I received the error alerts. In AGIS Server Manager the recycle times are set to defaults, midnight.

Other than wait to upgrade to 10.9.1 and hope the error goes away (I think it will...strictly based on my hunch), I am not sure what to do other than just deal with it. It is more of an occasional nuisance, fortunately my users have not noticed it.