Request Error Percent Alerts by the thousands since AGOL October 2025 Update

[DavidColey]

Not quite sure where to post this.... since the October ArcGIS Online update, all of my portal hosted feature layers that participate in my various 'By Reference' collaborations with our ArcGIS Online host have begun to throw constant Request Error Percent errors:

Logged At Component Id Code Level

10/24/25, 2:00:48.165 PM

624

9,000

severe

Source Message

Hosted/Street.FeatureServer

Hosted/Street/FeatureServer: null

Just over and over and over....

Probably associated with Open data or Hub, because otherwise there are no issues with the portal hosted layers in either Portal or Online,  and the hosted server site remains stable and normal.

Anyone else?

Thanks!

@DerekLaw 

[Eugene_Adkins]

Don't know the answer, but interesting that your post and this post were made in very close proximity to each other. They both, at the very least, share the same code ... though the code could be somewhat "generic."

[DavidColey]

yeah I saw that too, but thought that situation is referring to a geocode service issue and I'm not seeing that with my local geocoder services.

[DavidColey]

sometimes these things take a while to cycle through after updates, so we'll see.  Everything else is nominal

[DavidColey]

Here is an ArcGIS Monitor Chart showing the number of failed requests for our ArcGIS Server Host Site from 10/13/2025 - 10/27/2025:

Note there are essentially zero failed requests to the entire site prior to the 10/22/2025 ArcGIS Online update.  We changed nothing about our hosting site or our hosted feature layers (or servers, nor have we applied any patches to the 11.5 environment - everything is fully patched for 11.5 prior to the Online update).

This has to be related to the Online update and how the layers are collaborated across from our portal to ArcGIS Online by reference.

But again, there appear to be zero problems with the site, the hosted layers, the collaborations or anything with the servers themselves, e.g. memory use or network. 

Everything is working as expected.

I wouldn't even care except it's filling my logs 

@RussRoberts 

@JakeSkinner 

[DavidColey]

At least the Server logs are starting to show a different error message other than 'null'

Log Details

Logged At	10/28/25, 9:23:57.198 AM
Component Id	624
Code	9,000
Level	severe

Source	Hosted/SiteDevArea.FeatureServer
Message	Hosted/SiteDevArea/FeatureServer: Cannot invoke "com.esri.standardsql.query.Selection.getExpression()" because "<local6>.y" is null

Time	1761657837198
Type	SEVERE
User	null
Thread	1
Elapsed	0.0
Machine	mymachine.xyz.xyzx
Process	1340
RequestID	c41f7eb7-cab5-4ea4-8774-7874273fa204
MethodName	null

[DavidColey · Accepted solution]

Good day -

The repeating SEVERE-level log entries reporting a source errror:

Logged At Component Id Code Level

10/24/25, 2:00:48.165 PM

624

9,000

severe

..........

Source Message

Hosted/Street.FeatureServer

Hosted/Street/FeatureServer: null

Has been resolved.  I do not know why, but once I re-enabled the services directory on our Hosting site, the log entries ceased.  After a 15 minute wait, I re-disabled the services director on the Hosting site, and the log entries reporting a null source error have remained absent. 

So for us, the action re-enable re-disable the services directory solved this issue.

[DavidColey]

Interesting side note on this - our hosted server site is a 2-machine cluster.  whenever we have to restart the machines (after Windows updates, etc) the logging error re-appears until I go into server admin and re-enable / re-disable the services directory.  Once I do that, the SEVERE log entries cease.