TOLOGIX - Infrastructure LawGuide (ILG)

Subscribers Site - Performance Audit

Hi Morgan Maguire, CEO Morgan ,

Will has completed his audit on the subscriber's site. A disclaimer that developers often develop or use paid for tools to track and improve their own code over time but he assessed what he could. The issues may be also be related to dev environment and other hidden things being used right now for purposes on DevIT's end. Report below:


http://dev.infrastructurelawguide.com/SearchDocuments

Windows 10: Chrome 84

Scenario 1: Viewed different Tabs, switching between search and find

 
The two problem areas were;
1. On search Load it took upwards of 3 seconds to load.this is also true if for swapping between both tabs. 

 
2. Updating the list  took upwards of 6 seconds to receive and load content
 



The wait times are huge on the page. Receiving took 17 seconds to return everything and including 4 seconds of waiting in there for that.  I am not sure what is being loaded but whatever is in the 'Others' section of what's being loaded on the page takes up everything. I checked on other browsers and in some cases got even higher load times switching tabs upwards of 20-30 seconds. 

NOTE: There is normally entire teams dedicated to performance and performance improvements. I was not able to use the tool I had intended to for this but my hope is this will give some idea of whats happening. I also attached the .har from this. 

http://dev.infrastructurelawguide.com/SearchDocuments

Mac 10: Safari 13 

Scenario 2: Searching by keywords and using filters to refine results after showing


1. Problem area with loading again upwards of 3+ seconds to generate tab.

2. Second problem was after applying filters and results being generated.

Search and sorting just doesn't seem optimized. When using my own browser after the test I found it exhausting and tedious to look for documents. The loading and flashing every search/filter application plus the time it takes to give results was quick to wear down me visually. There is definite code optimization that needs to be done for results display and sorting. I'm experiencing upwards of 30 seconds loads in a few cases with larger results such as searching for 'Abbotsford Regional Hospital & Cancer Centre'. 

It is however much faster after revisiting results. I am not sure if this cookied after seeing something for the first time. 

Some other browsers I tried:

Checked on: 

Mac
  • Chrome 81
  • Safari 13

Windows 10
  • Edge Chromium 83
  • Edge 18
  • FireFox 72



If you have any questions, let us know. 

Savannah

Comments & Events

Morgan Maguire, CEO
Hi Savannah Mitchell, Project Manager at Industrial Savannah ,

We can discuss during the call, but did will perform an audit on the production environment as requested: http://www.infrastructurelawguide.com/login. This where we are having problems.

Morgan 
Savannah Mitchell, Project Manager at Industrial
Hi Morgan Maguire, CEO Morgan

Here is an updated performance audit on the subscribers side from Will: 

I performed a few scenarios for testing the loading of browse by topic and keyword search. The goal was to see if adding and removing filters, searching and other activities produced extended load times. I did some recordings of the testing to capture observed issues. 
 
Scenario 1: Purely Accessing the 3 main tabs in the search are produced, in my opinion, far longer loads than I would expect. 13 seconds be the worst offender with browse by topic.

NOTE: this load time varied between 4 and 13 seconds recorded, sometimes higher unrecorded upwards of 20 seconds.




Scenario 2: Browsing by topic, combining search with filters and changing them as I needed. Loading the initial search at its best was 15 seconds and at its worst was 1.25 minutes.



Scenario 3: Continued with browse by topic but did a deep dive into expanding all content shown, removing filters and other destructive and constructive actions with the data and search options. For some reason the time for removing filters via the clear button beside search was not recording properly. It was 4 minutes at the its worst to return with full list of browse topics when cleared.

Moving through searches once loaded was okay, but changing them and returning to browsing topics etc. takes 10+ seconds to reload.



From and outsider view it looks like content is being pulled ALL at once rather than limiting at the initial browse by topic or keyword only when a user searches.

Adding filters onto those searches isn't as bad of load times as expected given the rest of the experience using. Removing filters appears to basically break the site regardless of filter types or searchings.

I am not sure if we're caching the topics once its loaded but I did notice once results are there, the load times decrease a lot. My recommendation is to not display topic or keyword list there on tab click, but only after a user searches display exactly or near to what they're looking for rather than just a "here is everything list", which I doubt anyone is using anyways in that state.

Please note: Any changes or improvements should be discussed knowing that a full team is usually dedicated to the speed improvements of a site. I can only report what I'm seeing and not how to fix it and optimize code.

If you have any further questions, Morgan Maguire, CEO Morgan , perhaps we can set up a call to go through this in more detail. 

Savannah 
Morgan Maguire, CEO
Ok. Thanks Savannah Mitchell, Project Manager at Industrial Savannah . I'll add this to the agenda for tomorrow's meeting.

Also, were you able to pull the testing data in a easily readable format that can be analyzed by the development team?

Thanks,

Morgan 
Morgan Maguire, CEO
Ok. Great. Thanks Savannah Mitchell, Project Manager at Industrial Savannah .

Jitesh Dhuravala, DevIT Jitesh , please review the results above and we'll discuss during tomorrow's meeting.
 
Thanks,

Morgan 
Jitesh Dhuravala, DevIT
Hi Morgan Maguire, CEO Morgan ,

We are also facing performance issue to load data, we will look into it and check origin causes of issue. I have planned to publish in our environment with same data those are available in SME to make sure there is no server side issues when accessing data, after we will check technically. but sure we will overcome this issue.

Thanks,
Jitesh
Jitesh Dhuravala, DevIT
Hi Morgan Maguire, CEO Morgan and Savannah Mitchell, Project Manager at Industrial Savannah ,

I have configured same source code and database data over the DevIT server. You can check performance issue which Savannah has posted earlier in ILG. 

URL : http://islg.devitsandbox.com/Subscriber

Credentials : 
UserName : hiren.patel
Password : Admin@123

You can create user in subscriber side and check performance, We can discuss more in tomorrow call.

Thanks,
Jitesh
Morgan Maguire, CEO
Hi Jitesh Dhuravala, DevIT Jitesh ,

Before Savannah Mitchell, Project Manager at Industrial Savannah does any testing. What is the purpose of performing test in the this environment as opposed to http://www.infrastructurelawguide.com/login? We need improve the performance improved there, not in DevIT sanbox environment.

Also, there is pretty significant issue when attempting to view excerpts within the Subscriber site. I'm not sure if this has been reported through a to-do and user story yet, Naomi Joanis, UX Team Lead at Industrial Naomi , but we need to ensure this is resolved as high priority:



Thanks,

Morgan
Naomi Joanis, UX Team Lead at Industrial
Hi Morgan Maguire, CEO Morgan ,

I don't seem to have credentials for the sandbox environment that was discussed above and where you are experiencing that issue. This issue has come up before during UAT and was logged and I believe generally resolved (though I would need to go through this again more thoroughly to be certain).

When I go through the dev environment where I performed testing I am not experiencing this issue. Did you experience this outside of the sandbox environment?
William Roberts, QA Manager at Industrial
Morgan Maguire, CEO Morgan

I have not done the performance tests yet. Have the developers ran their own performance testing? What results have they seen in terms of improvement or areas with issues mentioned before?

William
Morgan Maguire, CEO
Hi William Roberts, QA Manager at Industrial William ,

As discussed last week, Jitesh Dhuravala, DevIT Jitesh has requested that selected elements that were experiencing problems on www.islg be tested against the results of the sandbox environment to see if there are material differences in the way the application is performing in both environments. 

This is a high priority item. We need to get these issues resolved in ww.ilg, so that we can start analyzing the subscriber side results of our SME tagging.

Morgan
Morgan Maguire, CEO
Great. Thanks Hiren Patel Hiren . We'll review and test the Subscriber site performance once we get the new ISLG launched later this month.

Thanks,

Morgan