Appending data to a Feature Service via arcpy or gp tool

Question

Hello,

My organization has recently upgraded to Enterprise 11.5, ArcGIS Pro 3.6. I noticed in a recent update that there is a new "Optimize for Feature Service" toggleable option when running an Append in ArcGIS Pro onto a Feature Service, either Reference or Hosted.

I know this option is newly available, and I understand that when checked, it will use the ArcGIS REST API instead of a direct arcpy Append.

We've noticed a lot of problems with Appends onto Feature Services since this option was released, both when that optimization option is checked and unchecked. Larger appends that previously might take 10/15 minutes are taking over an hour and are often failing.

See return from an append that ran over the weekend - this may be an outlier, as it's never this bad, but alas:

Start Time: Saturday, February 14, 2026 11:10:55 AM
250999 row(s) from (path/to/local/gdb) to (service_url)

Succeeded at Monday, February 16, 2026 6:19:30 PM (Elapsed Time: 55 hours 8 minutes 34 seconds)

"Optimize" was checked on this Append. There is one attribute rule on the sde feature class that this service is built from. I don't want to discount that as the cause of a bottleneck prohibiting real batching, but I've never had this level of sluggishness even with multiple attribute rules and editor tracking.

We've checked the health of our SDE (postgres) and can't detect any significant bottlenecks in the last month or so that would've affected this. It's only the append operations that are choking, other gp tools work just fine.

I mainly want to know if anyone has experienced anything similar recently, so we have a better understanding of what's going on.

Thank you!

RussellBrennan · Answer

Hi @SimonLux,

The optimize append mode on the GP tool will use the server Append operation to perform the append. There are some caveats to using this Append operation that need to be in place, you either need to be an admin on the service (all admins can append) or you need to have enabled the Append capability on the underlying feature service.

When we perform the optimized append we load data into a file geodatabase and upload the data to the server where the server takes over and loads the data directly into the data storage. This is typically a lot faster and has less chance of timing out that our other option which is to use the ApplyEdits operation.

In previous versions of Pro we would do this optimization for you if all the criteria was met for append so I am not sure if your issue is directly related to this new option. In our testing 250k features would never take this long with the optimization enabled, even without it enabled my expectation would be better performance than that. If possible think it would be worthwhile reaching out to Technical Support to get a repro case that we can take a look at.

You can view the web traffic using the built in Pro diagnostic monitor HTTP tab in Pro 3.6, this would allow you to identify if the slow append is using Append or ApplyEdits REST requests and could help shed some light on why this is so slow in your case.

SimonLux · Answer

Yeah we did catch that edit session/enable undo and have tried both ways. Both unoptimized and optimized are just taking extremely long, on both Hosted and Reference services. Not too helpful a response on my end, I apologize. But kinda at a loss.

DanPatterson · Answer

Appending data to a Feature Service via arcpy or g... - Esri Community

there are several cavaets as to whether it will be optimized..  This is worth checking since it is at the very bottom of the help topic.

When the tool is run in an existing edit session or the Enable Undo toggle button is turned on, the Optimize performance for feature services parameter will be ignored.