How to troubleshoot poor feature service performance?

795
10
Jump to solution
12-12-2016 09:40 AM
RoyceSimpson
Regular Contributor

I'm running a single cluster (3x 4 core servers) AGS 10.41 server system with web adapter in dmz and having serious performance issues on a feature service shared out to a web app builder app.  I can't share it to this post out as it is secured but I've tried disabling security with no positive effect.  Just panning and zooming the map can take upwards of 30+ seconds (there are only perhaps 30 features in the entire feature class).  If I turn the feature service visibility off in the TOC in my web app builder app, the map pans and zooms properly within a second or two. If I inspect the web traffic with the feature service I notice that there are upwards of 15 separate queries back to ArcGIS Server on every pan and zoom for that feature service.  Is that normal?  If I view the map service (not feature service) directly in the built in JavaScript viewer in AGS, it is very snappy.  Everything is pointing to something being wrong with this feature service but I can't figure out what it is.  I've viewed the AGS logs with no errors or warnings present.

Thanks for any help.

-Royce

0 Kudos
1 Solution

Accepted Solutions
RoyceSimpson
Regular Contributor

Answer:  Our shared file system where the config-store, output, jobs, etc... folders are located had "oplocks" enabled.  I had our sys admin disable that and now the services perform 10x faster.

View solution in original post

0 Kudos
10 Replies
JonathanQuinn
Esri Frequent Contributor

Is the data very dense and contain a lot of vertices?  If you have ArcMap on the same machine, does the data perform well in ArcMap at the same scale and in the same area?  You can take a look at this blog on how to see how many vertices your features have.

If you turn on verbose logging and resubmit the query request, (you'll probably want to do this during downtime so you can single out the request), you can take a look at each operation that makes up the request.  You'll see things like symbolizing, data access, etc.  From there, you can narrow down the problem.

0 Kudos
RoyceSimpson
Regular Contributor

It's point data and only about 30 points in the whole feature class.  Archiving is turned on however, so I'm sure there's more overhead behind the scenes, but I can't imagine there's a ton.  I do have verbose logging on and haven't seen anything stand out.  In Arcmap, the data draws extremely quickly.  As fast as you'd expect for 30 points in a feature class.

0 Kudos
JonathanQuinn
Esri Frequent Contributor

If you were to run a query of 1=1 at the Rest endpoint against the layer in both the map and feature service endpoints, do you see the same performance?

0 Kudos
RoyceSimpson
Regular Contributor

Thanks for working with me Jonathan.

I ran a 1=1 query on the map service (layer 0) and it returned all 37 in less than a second, same result several times.

The same query via feature service layer zero took 2 to 4 seconds each time.

I've noticed in Chrome Dev tools that a bunch of queries are fired off to AGS from the client... like 15... all are in a wait state, then about 30 or more seconds later, they trickle in and finish.

0 Kudos
RoyceSimpson
Regular Contributor

Also, may be worth noting that I have a max of 5 instances per machine allocated for that service for a total of 15.

0 Kudos
JonathanQuinn
Esri Frequent Contributor

I see... can you describe the data and your setup?  What version of ArcGIS Server?  What DBMS is the data stored in?  How many fields are there in the data?

0 Kudos
RoyceSimpson
Regular Contributor

Server 10.41.  Oracle 11g.  12 fields on the data.  

Jonathan, are you an esri employee?  I can create a temp login for you so you can see the feature service.

-r

0 Kudos
JonathanQuinn
Esri Frequent Contributor

Hey Royce, yes I am.  I'll send you a message.  I won't be able to do much troubleshooting from my end, so it may be best to reach out to Technical Support.

0 Kudos
RoyceSimpson
Regular Contributor

Hey Jonathan.  I'm just going to log a ticket on this.  I'll keep this thread up to date as things progress.

-r

0 Kudos