Portal Running Very Slow

[DeanHowell1]

Our production portal in AWS is running extremely slow and was wondering if people have had that is issue and can give me some hints in how to fix?

I know this is a very open question, but I am at a dead end and not sure what to try.

Things I have tried:

• Restart services
• Restarting servers

[ReeseFacendini]

What is the instance type of the machine (i.e. t2.large)?  Do you have the other pieces of ArcGIS Enterprise on that machine as well or are they distributed?

[DeanHowell1]

Hello @ReeseFacendini we have each part of the enterprise on different servers but in the main they are either r5a.2xlarge or m5a.xlarge instances. 

[jcarlson]

I'd also check: on whichever machine Server is running on, how many instances of ArcSOC.exe are there in Task Manager. Assuming you're on a Windows machine, that is.

[DeanHowell1]

Thanks @jcarlson I appreciate your suggestion. There are a lot of ArcSOC services running on some of the machines so that might be part of the issue.

[jcarlson]

A colleague of ours had similar issues. A lot of their "minor" services were being published as dedicated instances, which ended up hogging tons of RAM. Moving some of the less-important services to the shared instance pool will really help, if that's the case.

Edit: If the services don't require an enterprise geodatabase's functionality, you might also consider moving your services to hosted layers. Since 10.8, Hosted Image Layers were added, so the lack of "fancy" rendering in hosted layers doesn't necessarily have to be a dealbreaker. 

[DeanHowell1]

Thanks Josh, yes I think that is worth looking at. There are a lot of hosted layers but also a lot not hosted 🙂

[DavidHoy]

Hi Dean,

is it the Portal pages generally - or when previewing specific services or the web maps/scenes used by the services?

Have you examined the ArcGIS Server logs using System Log Parser to identify which services may be the chief culprits? 

[DeanHowell1 · Accepted solution]

Thanks David, it was all pages in the enterprise but what we discovered is that AWS had rotated IP addresses, so Portal was spending a lot of time trying to authenticate every action. We had to update the DNS records and speed returned to normal.

[SubodhWMK]

Could yo uplease elaborate more what DNS records did you updated ? We are going through similar issues where our portal is extremely slow. Also, our applications are having to reauthorise every half an hour, when inspecting web traffic, app is looking to get access token to authenticate user.