# Connection Failures and system too slow

**URL:** <https://forum.inductiveautomation.com/t/connection-failures-and-system-too-slow/64855>\
**Category:** Ignition\
**Tags:** vision\
**Created:** [October 5, 2022, 4:51pm UTC](https://forum.inductiveautomation.com/t/connection-failures-and-system-too-slow/64855 "2022-10-05T16:51:45Z")\
**Posts on this page:** 9\
**Page:** 1

<div class="post-metadata">

**Author:** ![Reyes](https://avatars.discourse-cdn.com/v4/letter/r/ccd318/32.png) [@Reyes](https://forum.inductiveautomation.com/u/Reyes)\
**Post date:** [October 5, 2022, 4:51pm UTC](https://forum.inductiveautomation.com/t/connection-failures-and-system-too-slow/64855/1 "2022-10-05T16:51:46Z")

</div>

Hello forum,

I'm having some different issues with my SCADA system.

**Clients:**  
All the clients are running too slow. The system does not respond appropriately when we try to navigate, select, or change something on any screen. Also, everything changes to red every minute, and it returns to normal.

**Designer**  
To do any change on my designer, the system it's too slow, and I can see every minute a little pop-up message saying "Connection Lost" and at the same time another one "Connection Restored." All day the system it's doing that.

I increased my memory to initmemory=2048 maxmemory=8192

The limit of concurrent connections between the client and the gateway. _"Connection Concurrency"_ option in Project Properties-\>Vision-\>Timing: IT'S DISABLED.

Any advice? I'll appreciate it.

Note: Project Ignition 7.9.16 was created years ago. I just started with this company, and they have a lot of old screens not in use (on red). I'm thinking of cleaning all  
 ![ErrorIT](https://us1.discourse-cdn.com/inductiveautomation/original/3X/8/0/8001e68b22368a8702578c79897213e0fba45fb5.png)  
the SCADA and starting over.

Thank you so much forum! 😃

---

<div class="post-metadata">

**Author:** ![YF129701](https://sea1.discourse-cdn.com/inductiveautomation/user_avatar/forum.inductiveautomation.com/yf129701/32/60498_2.png) [@YF129701](https://forum.inductiveautomation.com/u/YF129701)\
**Post date:** [October 5, 2022, 5:05pm UTC](https://forum.inductiveautomation.com/t/connection-failures-and-system-too-slow/64855/2 "2022-10-05T17:05:02Z")

</div>

A few obvious things you can start with:

- What is the performance of your gateway look like? How much memory/CPU is being used? Did you restart the gateway after you allocated the additional memory?
- What do your clients looks like in terms of hardware? Are they running low on memory/processing power?

From my little Vision experience, it seems like a slow Vision app usually means the client lacks key hardware resources and the fact that it disconnects/connects to the gateway might be a gateway resource issue. At least those are the two most obvious, glaring things to check.

---

<div class="post-metadata">

**Author:** ![pturmel](https://sea1.discourse-cdn.com/inductiveautomation/user_avatar/forum.inductiveautomation.com/pturmel/32/3295_2.png) [@pturmel](https://forum.inductiveautomation.com/u/pturmel)\
**Post date:** [October 5, 2022, 5:08pm UTC](https://forum.inductiveautomation.com/t/connection-failures-and-system-too-slow/64855/3 "2022-10-05T17:08:31Z")

</div>

So many possibilities. Share more about your gateway. Performance page from the status section. How many tags does your system have? How many are OPC tags? How fast are they polling? How is your database set up? Shared server? Virtual servers?

What shows in your wrapper log file? "Clock drift" warnings? Maybe you need java GC logging to look for pathological memory problems?

{ ps: initial and max memory should be the same on production systems. Java never gives memory back to the OS, so it should claim it all up front. }

---

<div class="post-metadata">

**Author:** ![Reyes](https://avatars.discourse-cdn.com/v4/letter/r/ccd318/32.png) [@Reyes](https://forum.inductiveautomation.com/u/Reyes)\
**Post date:** [October 5, 2022, 5:15pm UTC](https://forum.inductiveautomation.com/t/connection-failures-and-system-too-slow/64855/4 "2022-10-05T17:15:51Z")

</div>

![Diagnostics](https://us1.discourse-cdn.com/inductiveautomation/original/3X/d/4/d4bbe424d4ad4e6390be311c2b7469e2600045d0.jpeg)

---

<div class="post-metadata">

**Author:** ![Reyes](https://avatars.discourse-cdn.com/v4/letter/r/ccd318/32.png) [@Reyes](https://forum.inductiveautomation.com/u/Reyes)\
**Post date:** [October 5, 2022, 5:16pm UTC](https://forum.inductiveautomation.com/t/connection-failures-and-system-too-slow/64855/5 "2022-10-05T17:16:07Z")

</div>

![Performance](https://us1.discourse-cdn.com/inductiveautomation/original/3X/1/3/136b361c31b8e63dc2479c395bd23b27442516eb.jpeg)

---

<div class="post-metadata">

**Author:** ![pturmel](https://sea1.discourse-cdn.com/inductiveautomation/user_avatar/forum.inductiveautomation.com/pturmel/32/3295_2.png) [@pturmel](https://forum.inductiveautomation.com/u/pturmel)\
**Post date:** [October 5, 2022, 5:20pm UTC](https://forum.inductiveautomation.com/t/connection-failures-and-system-too-slow/64855/6 "2022-10-05T17:20:01Z")

</div>

Your gateway's garbage collection algorithm choice is terrible. Remove those settings from ignition.conf. You want it to use ~~the default~~ G1GC. Also set the target pause for G1GC to ~100ms.

Your client chart shows a very light load. Very unlikely to be a client problem.

---

<div class="post-metadata">

**Author:** ![Reyes](https://avatars.discourse-cdn.com/v4/letter/r/ccd318/32.png) [@Reyes](https://forum.inductiveautomation.com/u/Reyes)\
**Post date:** [October 5, 2022, 5:38pm UTC](https://forum.inductiveautomation.com/t/connection-failures-and-system-too-slow/64855/7 "2022-10-05T17:38:11Z")

</div>

![IGNITION](https://us1.discourse-cdn.com/inductiveautomation/original/3X/7/6/767aee3223de8b85faf2de09b346fcd3acfbcca7.jpeg)  
pturmel,

Thank you for your help. How can I remove those settings from ignition.config? See the attached pictures of my ignition.conf

---

<div class="post-metadata">

**Author:** ![pturmel](https://sea1.discourse-cdn.com/inductiveautomation/user_avatar/forum.inductiveautomation.com/pturmel/32/3295_2.png) [@pturmel](https://forum.inductiveautomation.com/u/pturmel)\
**Post date:** [October 5, 2022, 5:41pm UTC](https://forum.inductiveautomation.com/t/connection-failures-and-system-too-slow/64855/8 "2022-10-05T17:41:01Z")

</div>

See this post:

> [@Migrating to Java G1GC garbage collection](https://forum.inductiveautomation.com/t/migrating-to-java-g1gc-garbage-collection/18638):
>
> Reading up on [G1GC in forum posts](http://forum.inductiveautomation.com/search?q=G1GC), I’ve gathered: Clients Add jvm-args="-XX:+UseG1GC" to the end of the target string in the native client launcher shortcuts to use G1GC under Java 7 (from [this post](http://forum.inductiveautomation.com/t/openjdk-and-linux/18553/6) by @pturmel). On the other hand, I’m thinking this is not required as long as clients are running on Java 8+. See next point. Consider limiting Java versions to 8+ in gateway settings as G1GC is the default in Java 8 per [this post](http://forum.inductiveautomation.com/t/gateway-java-question-7-7-2/10402/6) by @pturmell. Also note [G1GC is not fully supported prior to JDK7 r…](http://www.oracle.com/webfolder/technetwork/tutorials/obe/java/G1GettingStarted/index.html)

Note that you must explicitly set G1GC in java 8. It didn't become the default until java 9.

---

<div class="post-metadata">

**Author:** ![pturmel](https://sea1.discourse-cdn.com/inductiveautomation/user_avatar/forum.inductiveautomation.com/pturmel/32/3295_2.png) [@pturmel](https://forum.inductiveautomation.com/u/pturmel)\
**Post date:** [October 5, 2022, 6:55pm UTC](https://forum.inductiveautomation.com/t/connection-failures-and-system-too-slow/64855/10 "2022-10-05T18:55:37Z")

</div>

You are missing the line that actually activates G1GC.  
You are deleting unrelated parameters (10-14). Probably should leave those there.  
You will end up with two sets of memory limits. Don't do that, either.
