# OPC-UA performance issues after upgrading to 7.9.10 **URL:** \ **Category:** Ignition\ **Tags:** ignition79\ **Created:** [June 11, 2019, 8:02pm UTC](https://forum.inductiveautomation.com/t/opc-ua-performance-issues-after-upgrading-to-7-9-10/25744 "2019-06-11T20:02:46Z")\ **Posts on this page:** 6\ **Page:** 1
**Author:** ![Peter.Dmitriev](https://avatars.discourse-cdn.com/v4/letter/p/7ea924/32.png) [@Peter.Dmitriev](https://forum.inductiveautomation.com/u/Peter.Dmitriev)\ **Post date:** [June 11, 2019, 8:02pm UTC](https://forum.inductiveautomation.com/t/opc-ua-performance-issues-after-upgrading-to-7-9-10/25744/1 "2019-06-11T20:02:46Z")
Hello, We recently upgraded from 7.8.2 to 7.9.10. After that opc-ua bound buttons and indicators started flickering at random. Looking into tag browser I can see the same behavior. Gateway performance (CPU & Memory) appears to be normal. Scan class for all tags is the default 1,000ms. Any suggestions on performance tuning, etc.? ![image](https://us1.discourse-cdn.com/inductiveautomation/original/2X/f/f4e236cab70a5ba1fd0ce8147cc8278c1b4061e3.png) ![image](https://us1.discourse-cdn.com/inductiveautomation/original/2X/7/7a24241652ed5f73cd8533e010b3e0a5ec89bb50.png) Thank you, Peter. ---
**Author:** ![Kevin.Herron](https://sea1.discourse-cdn.com/inductiveautomation/user_avatar/forum.inductiveautomation.com/kevin.herron/32/7259_2.png) [@Kevin.Herron](https://forum.inductiveautomation.com/u/Kevin.Herron)\ **Post date:** [June 11, 2019, 8:33pm UTC](https://forum.inductiveautomation.com/t/opc-ua-performance-issues-after-upgrading-to-7-9-10/25744/2 "2019-06-11T20:33:31Z")
Which driver are these tags from? ---
**Author:** ![Peter.Dmitriev](https://avatars.discourse-cdn.com/v4/letter/p/7ea924/32.png) [@Peter.Dmitriev](https://forum.inductiveautomation.com/u/Peter.Dmitriev)\ **Post date:** [June 11, 2019, 8:37pm UTC](https://forum.inductiveautomation.com/t/opc-ua-performance-issues-after-upgrading-to-7-9-10/25744/3 "2019-06-11T20:37:37Z")
They are from **Allen-Bradley Logix Driver** (processors with firmware 21+). This is the only device driver we use on this gateway. ---
**Author:** ![Kevin.Herron](https://sea1.discourse-cdn.com/inductiveautomation/user_avatar/forum.inductiveautomation.com/kevin.herron/32/7259_2.png) [@Kevin.Herron](https://forum.inductiveautomation.com/u/Kevin.Herron)\ **Post date:** [June 11, 2019, 8:39pm UTC](https://forum.inductiveautomation.com/t/opc-ua-performance-issues-after-upgrading-to-7-9-10/25744/4 "2019-06-11T20:39:29Z")
Ok. There were some performance regressions between 7.9.8 and 7.9.10. You should upgrade to 7.9.11. Additionally, check the concurrent requests and CIP connection size settings. I think there were some different defaults being used back in 7.8.x. ---
**Author:** ![Peter.Dmitriev](https://avatars.discourse-cdn.com/v4/letter/p/7ea924/32.png) [@Peter.Dmitriev](https://forum.inductiveautomation.com/u/Peter.Dmitriev)\ **Post date:** [June 11, 2019, 9:21pm UTC](https://forum.inductiveautomation.com/t/opc-ua-performance-issues-after-upgrading-to-7-9-10/25744/5 "2019-06-11T21:21:05Z")
Kevin, CIP Connection Size is 500. Just checked my other 7.7.10 and it’s the same there. Not sure what was it in 7.8.2. Could increasing or decreasing that setting help? I also just discovered this: ![image](https://us1.discourse-cdn.com/inductiveautomation/original/2X/2/25a90a1c16e8c0a65e0392a21e73c05d7262bc9c.png) I don’t have any 100ms scan classes and all 3 transaction groups are @1sec. What’s trying to read tags every 100ms then? I’ll start looking into upgrading to 7.9.11 but it’s a significant effort due to downtime coordination, etc. (no redundancy on that gateway yet). ---
**Author:** ![Kevin.Herron](https://sea1.discourse-cdn.com/inductiveautomation/user_avatar/forum.inductiveautomation.com/kevin.herron/32/7259_2.png) [@Kevin.Herron](https://forum.inductiveautomation.com/u/Kevin.Herron)\ **Post date:** [June 11, 2019, 9:24pm UTC](https://forum.inductiveautomation.com/t/opc-ua-performance-issues-after-upgrading-to-7-9-10/25744/6 "2019-06-11T21:24:31Z")
Check your scan classes and transaction groups again. Don’t trust scan class names, I’ve seen the supposed rate in the name not match the actual configuration many times. Increasing the CIP connection size and concurrent requests may help but you need to experiment with it to find out. It only helps if the PLC comms aren’t already overloaded.