Speed

We are in the process of moving to the cloud and using BIRST as our primary reporting tool. Currently, on our training (test) site, CSD Analytics is incredibly slow. It takes at least five minutes for the initial screen to load and then another five to navigate to Visualizer and Dashboard. When creating a new report or a new dashboard, sometimes it's so slow that it literally times itself out and we get a "Data is processing come back later" or "Whoops something is missing" error message. Is this normal for BIRST? I've never worked with a BI tool that processed data or changes so slowly.

Find more posts tagged with

Comments

Legacy Contributor

Unfortunately we have been experiencing the same all the time. And it gets worst when adding or modifying filters in a Dashboard, you can type a name of a filter in the search box and still what you type is not displayed, go grab a coffee and wait another 3 or 4 minutes till you see what you typed starting to show on the search box then wait another 5 minutes till it is pulled.

I have worked on multiple Birst releases and still the same issue.

hopefully one day Infor/Birst will understand that this is a global issue that needs to be addressed.

andy

Hearing this is a little concerning for us as we begin the move from Single Tenant to Multi Tenant.

Have there been incidents opened regarded these latency issues? Was there any explanation given?

karin-drinkwaard

Hi,

We are in the cloud and use Birst now for more then a year.

We experienced the same and were told this had to do with the fact that the initial screen has to much data on it.

Therefor we have split the start dashboard in 5 "blocks" with a total of 13 (K)PI's

f.e. if you click on the "block" that is called Backlog - you go to the backlog dashboard that contain a total of 9 (K)Pi's

The initial dashboard takes around 20 seconds to load, but the others are then appearing within 5 seconds.

Hope this helps.

Peter Vermeulen

Yep, we have the same. We looked at replacing our current BI tool (Tableau) with Birst since it comes together in our Cloudsuite, so we have it up and running for evaluation but it is giving the above issues so it keeps being in the exploration phase but no intention yet to use it for all of our BI needs.

Legacy Contributor

We have experienced the same and we'll not recommend BIRST as your main reporting tool.

We were wrongly advised and over sold on BIRST and would have been better off spending the time bringing the data locally and developing reports in PowerBi, SSRS or even Excel.

Other things that are not working for us are:

Fiscal year reporting. neither CSD nor Birst support fiscal year reporting that does not start on Jan 1st
Custom content. Since Birst reports are overwritten every time there is a new update, maintaining custom content becomes a hassle.
Inadequate for adhoc reporting or to extract master data. i.e., extracting ARSC, APSV, ICSP and don't even consider ICSW
Reporting between sales orders, POs and WTs is not supported, even though the ties exist in CSD, Birst model does not include them and is not possible to bring PO data for tied sales orders.

I've heard of these other tools that might be good to evaluate: Phocas, Hyperion, Acumen, Infor ETL

We’ve heard thru our Partner that Infor is investing heavily on improving the tool and are hopeful the tool does improve.

ryan-staley

We've had the same experience and are now heading the data warehousing / Power Bi route. We have found the tool slow, unintuitive, and the documentation often outdated and cursory. It just isn't mature enough to use as a primary reporting tool at this point.

Legacy Contributor

All these observations are completely valid but I think its important to clarify that this is not a BIRST specific issue but related to the data model that has been developed for BIRST. CSF Analytics and HR Talent Analytics (which I think most people here are referencing) have a lot of meta data and loads transactional data. If you were to build out a similar model in other tools you would likely encounter similar struggles.

For those wishing to improve the viability of the pre-delivered content be sure youre reaching out to Infor directly as there are some key items that can help with performance.

If you are interested in designing your own trimmed down data model, BIRST is just as capable as some other tools listed here. its just a matter of understanding the source data, which will be a struggle for many regardless of the end repository. Infor is learning about the gaps in their data models, as you would, were you to create a custom solution in any tool. The bonus here is you get monthly releases with new content and performance improvements.

Again, not invalidating the struggles, they are real, but be sure to weigh that against what you will inherit on other platforms and the work effort that goes into getting a usable data model.

Diane McElhiney

@mnye Well stated. And I'd like to remind those CSD Analytics customers facing challenges to communicate to the Customer Success Team and they in turn can coordinate a discussion with a Technical Success and/or Technical Account Manager from Birst.

Legacy Contributor

Hi Diane, We have opened so many tickets and have meetings with several CSM and every time we spend many hours with the Technical support team and Birst engineers and still this issue is not resolved and we are still facing that slowness. We keep getting promises that the issue will be resolved with the next release but it never resolved.

Thanks

Ahmed

Legacy Contributor

I'd like to add to what @mnye said. I have seen custom-built Birst spaces for CSD and SXe and M3 and others that perform very well. Some points to consider:

1. Birst is not simply a BI tool in the way that Tableau and Power BI are. Birst is a data warehousing platform and a BI tool. It extracts data from a very broad range of sources, transforms the data however necessary, and loads the data into a star-schema data warehouse. I have built data warehouses manually using various ETL tools, and Birst's capabilities in this area are top notch. It is much easier to build a data warehouse using Birst than any other tool I've tried. In this it is comparable to SSIS and Data Factory and Azure, not to Tableau. But in my opinion it's much easier and faster to go-live than SSIS/Data Factory, etc.

2. You mention "Data is processing come back later" - that error is displayed when the ETL process is occurring and the data is being updated in the data warehouse. Birst has a feature called "Always On" which allows end-users to continue viewing dashboards, etc. while data is being processed. My understanding is that it is not turned on in the CSD Analytics solution (and other pre-built solutions) due to the multi-space complexity of those solutions. I have seen it used very effectively in more straightforward situations.

3. That "Whoops something is missing" error is maddening and needs to go away. If you are experiencing that repeatedly you should definitely open a support ticket with Infor. I was seeing that every minute or two when I was trying to do anything in Visualizer and the Dashboard and Infor support was able to do something (I don't know what) that resolved the issue completely for me.

4. In my opinion, the pre-built analytics solutions were implemented in a particular way that was intended to provide broadly requested reports to a broad audience and to optimize the solution to allow Infor to push out updates to all customers using the content. I could write pages and pages about what is not effective in the pre-built content, why they're so slow, why they're so frustrating to customize, etc. But those things are not indictments of Birst itself. I have built CSD, CSF, M3, etc. spaces that perform much better than the pre-built content, but they are built to the specific needs of specific companies and are not intended to be comprehensive, all-in-one solutions that instantly include reporting in all business areas.

5. A lot of the issues in the Birst solutions for the CloudSuite ERP's are due to the fact that the extract is from the Infor Data Lake. The Data Lake is not a database, per se. There are no tables holding all the data. Data is held in basically json files, and typically every single change to every single record is in the data lake. Due to this, extracting data from the data lake will be slow whether it's Birst doing the extract or some other tool. Also, since the data lake typically has every single record change, this adds complexity and time to the ETL process because it needs to be reduced to only the most recent version of the record.. A Birst solution built off of an SQL database typically completes the ETL process much faster than the same Birst solution built off of the data lake.

6. To @aelsaie , if you are experiencing those kinds of issues when adding or modifying filters you should definitely reach out to support. It sounds infuriating, especially if you've seen it across multiple Birst releases. I have never seen that happen in a healthy, well-setup space. If Infor Support doesn't resolve it for you then you may want to consider reaching out to Birst Services or to a Birst Partner for expert advice on how to remedy the situation.

7. To @jim-lopez , Birst absolutely supports fully custom fiscal year reporting that starts on any day you would like. However, it has not been implemented that way in the CSD Analytics solution. Maintaining custom content should not be a hassle and it will not get wiped out if it is created in specific collections/folders. Furthermore, updates to the Analytics solutions are "pull" updates now, not "push" updates. Your solution will only be updated when you choose and the updates do not need to be applied if there is nothing new that you need. I recommend reviewing the notes prior to applying any updates so that you can decide whether there is anything new that you were looking for. I don't understand your comment where you state, "Inadequate for adhoc reporting or to extract master data. i.e., extracting ARSC, APSV, ICSP and don't even consider ICSW." I have used each of these data lake tables in Birst solutions and had no problems with the extraction. I agree that reporting between sales orders, POs and WTs should be available in the pre-built solution. Have you put in a request for this capability to be added?

My suggestion to all is to not throw the baby out with the bathwater. It is unfortunate that you are all experiencing such poor performance. I won't try to suggest that Birst is perfect, and I won't pretend that there isn't a long list of "technical debt" that I wish the Birst software engineers would get around to dealing with, but it's still the only comprehensive tool of its kind that can extract data from the Infor data lake, transform it into data warehouse in a schema intended for reporting, and deliver up the information on web-responsive dashboards that look good on every screen, all in a single, browser-based tool.

I built an entire solution for a company, from initial extract all the way to several dashboards, in about 12 hours over 2 days earlier this week. The solution includes sales, open orders, invoices, and daily inventory. It's not a tiny data set - 5 years of data with about 75M rows in the sales fact table. I don't believe there is any other data warehousing and BI platform that I could do the same on, other than Birst.

Legacy Contributor

Similar to Cindy's comments, we have built a repeatable solution for CSI/Syteline and ISM that provides more granular reporting vs. the the standard content. We never experience the noted speed issues. It's been stated in other ways already, but I would not want users to confuse the capabilities of Birst as a tool versus data modeling deployed in a particular Birst space or collections of spaces.

Legacy Contributor

Good all and thank you all for voicing your concerns. We are continually improving the product with each update release, incorporating suggestions from our customers. There were performance updates in February 2022 and more to come in June 2022.

Enhancements released in February include:

Birst now only downloads the color palettes used in a dashboard and provides tools for managing them. Resulting in faster UX load times.
Subject area navigation has no longer downloads the full subject area to searching and incremental download resulting in faster UX load times.

Enhancements coming in June include:

Server-side caching even faster render times
Analysis of the memory used in the repository and subject area data structures was performed
Decreased the size of the repository in memory by 7-10x
Decreased the size of the subject area in memory by 3-5x
Further decreasing the memory used in SMIWeb
Moving to Java 11 (from Java 8) resulting in significant memory savings. Java 8 stores all Strings in ‘char’ arrays (2 bytes per character) vs Java 11 tries to store Strings in ‘byte’ arrays (1 byte per character) where possible = faster runtime
Changing the number of upload (extract) threads (Controlled Availability)
Faster overall extract times (and thus ETL/orchestration) since more can be done in parallel
Compass Data Lake Connector, complex query support and deduped records on receipt into Birst
Birst generates physical SQL based upon the logical BQL query moving to SQL Server 2019 will improve overall query performance.
Working on rolling this out to all customers as we speak – target completion end of 2022

It is our top priority to make sure the solutions we offer meet your needs. I am happy to announce that the updates coming in June will remedy many of the concerns I have seen throughout this thread. The changes are being beta tested and have resulted SUBSTANCIAL improvements and we look forward to rolling those out to you. We commit to hosting an information webinar closer to that time.

Legacy Contributor

[quote userid="60674" url="~/birst-user-community/f/birst---general-discussions/28217/speed/75468"]

Decreased the size of the repository in memory by 7-10x
Decreased the size of the subject area in memory by 3-5x

those are the 2 im looking forward to the most. Ive heard some of the spaces are going away as well which should be a huge boost in both rendering of the visualizations, filters, ect and orchestration.

Legacy Contributor

Interesting conversations and I recognize the trouble and frustration that people got. I'm working at a company that uses M3 Single Tenant Cloud since mid 2019 together with Birst Cloud.

We started with M3 Analytics in Birst, but gave up. To slow and hard to customize.

Went to Birst Professional, slow and useless.

Tried Birst Enterprise Live Access, to slow and hard to customize.

Went to Birst Enterprise ADR, great when we finally understood how it worked.

Since then, we have gone to Birst (ADR Enterprise ) Networked BI and it's great.

Our next step is to change to change from multi-level to single-level hierarchies, incremental load and surrogate keys.

The documentation is still quite messy with Classic Birst and Pronto mixed. It's been hard to find best practices as well. It has taken years for us to find an architecture that we are happy with.

Petter

Lisa Callahan

@yasmin-barnes I know the appliance releases are a little behind the cloud releases. When will the above improvements be available for us?

Distribution Management Early Adopter 20241016.pdf

Legacy Contributor

You may want to consider engaging an Infor Birst Technical Account Manager (TAM) or engage someone from a partner (like myself!) to provide a few hours of mentoring each month. It would help you figure out the appropriate architecture and those best practices that apply in your specific scenario.

Legacy Contributor

Hi Petter

I appreciate your incite. 2 questions, given the amount of time, elbow grease (and probably money) what advantages has Birst given you that has made it worthwhile staying with Birst vs switching to another platform? Do you think you will every move back to the cloud, even if you keep your custom models? Thx

0911060553270289.pdf

Legacy Contributor

Before we implemented Birst in 2018-2019, we used Qlik View. Back then Birst was in the transition between Flash and Java (Classic/Pronto). It lacked a lot of functionallity. It has become better and better over the years:

The things I like with Birst is:

1. ADR modelling is really fast and powerful once you understand it. Build large galaxy schemas with 20 fact tables and 40 dimension tables has been no problem.

2. Network BI is a really good and simple way to distribute data.

Overall: Integrated (data warehouse and BI tool) cloud based solution that is simple to manage with small resources.

The Java web UI is still quite buggy/messy/slow compared to Qlik/Tableu/PowerBi installed applications. But it's getting better for every release and is easy to distribute since it's entirely cloud based.

We are running Birst in the Cloud right now, but are actually in the middle of a project migrating from M3 Single Tenant Cloud to M3 Multi Tenant where Birst is integrated via Infor OS. We will start converting our current enviroment next week, pulling all data from Data Lake.Then start using Incremental load as well since it's mandatory when pulling data from Data Lake.

We will take another look att M3 Analytics to see if it's worth using during the project.

Diane McElhiney

@ahmed-tahaelsaie

Can you please connect to message me? I'd like to research you challenges to speak with my colleagues to follow-up.

Legacy Contributor

Thanks, I looked at Birst when Infor first purchased it and thought it looked great. Since then I worked primarily with Power BI. We are just starting to roll out Birst for our customers (SyteLine ERP). It has been a rocky start, but I am glad to hear that although the product still has some growing pains to get through we can make it work in the long run.

Marie Baker

SPEED - I have learned how to make this work, and I want to share some steps that you can use to Improve Speed. Let me tell you why i kept moving forward with Birst, is because what it does best is bring back very accurate data. I have had to defend my numbers often and it has never let me down. But I have felt the "SPEED Issues" and I learned how to make it work. So I need to share what i've learned to help you: (1) Signing into Birst shouldn't take a long time. If it does, then it has something to do with internet speed but not with Birst. Some Countries ping different Birst servers and some countries have quicker internet speeds then others. (2) Birst is a Cloud package and the Birst Server(s) will Time you out to free up availability for other users if you have been inactive. In Birst, You won't realize you've been Timed out, because the last screen you used is still displayed on your screen_. So you might try to apply Filters, and it spins forever and you can't tell if it's doing something, or doing absolutely nothing. Other times you get the Whoops screen TIP: Save Often. TIP Close down the Browser if you are spinning or have been inactive for 20-30 minutes TIP: Yes I said. Close The Browser. Now Don't stop reading. Closing the Browser clears the Birst cache . Since I didn't want to close all of my other Applications that are running on my Chrome browser, I installed a second ( Fire Fox ) Browser and then I launch the FireFox Browser and I use it only for running Birst . That way i can end "X" out of the Fire Fox Browser any time Birst gets slow, and I can sign back in right away and when i do, I get a new Internet connection and new server connection and my Speed is back. (4) . If Data Extractions are running, you shouldn't be trying to Edit your Dashboards. Tthat's when Birst is extremely slow .Make sure to find out when your Refresh jobs are Scheduled to run. Avoid using Birst during those times. Also i have seen if you are making changes to a Dashboard while a Data Refresh is running it can and will corrupt the Filters on your Dashboards. (5) SPEED Tip: When Building Dashboards/Reports : Add as many KPI's and graphs as you wish . But you MUST add a Filter - that will limit Initial data that will be be displayed. Add a Date Filter on the Dashboard, such as "Weeks ago" = 1 Limit the initial Data. I work with millions of records and calculate hundreds of KPI's and produce hundreds of detail spreadsheets records for over 273 Cost Centers (Daily/Weekly/Monthly reports). After a Dashboard with limited data populates quickly your Users will like it. Then they are gong to Filter the Data to meet their own needs. So you give them a Tip of the Iceberg luring them in. . (6) Add some Text Boxes : "User Tips: to explain to the Users how to use the Dashboard. Tell them certain Filters are Mandatory Embed a Filter on the Dashboard that they see right away so they choose ex. Product and a Date Range and they will then limit data retrieval to meet their needs. (7) Add a Text Box : with Warnings: Warn that Selecting ALL will take a few minutes to Retrieve Data. Level setting your users will give them the patience needed. They will love to Export their own Reports Clicking on URL's on Columns in a Birst, Query to see the Source of the Transaction when you use the Report Level Drilling. Use your Linking to click on a Graph bar, to launch a supporting detail spreadsheet. It's also very easy to add Notification jobs to Distribute via email spreadsheet reports . We distribute our Severity One Call Report to over 300 employees. I have also used Tableau and Power BI ,and Cognos, and Corvu and all perform the same functions, and all have their quirks Once you resolve to learning something new, and dedicate yourself to working with it, you will love your BI software, whatever package you choose. Birst works for me. It actually becomes pretty easy to use for your non technical End Users. So good Luck - Hope some of these Tips work for you.

Legacy Contributor

Excellent advice, @marie-baker . I'd like to challenge your point 2) a little bit, however, because that may not be the same for everyone. People may not have their browser's set to automatically clear the cache when their browser closes. And it may not be necessary anyway. If you believe your browser is behaving strangely and has possibly lost its connection to the Birst server, try the following:

1. Refresh the browser tab. If you are logged in to Birst via Mingle you will have to get back into Birst, which is a pain in the neck, I'll admit. But if you are logged in to Birst directly it should just refresh your connection. If your login session has expired it will bring you back to the login screen.

2. If that doesn't resolve it, clear your browser's cache. YOU DON'T NEED TO CLEAR EVERYTHING FOR ALL TIME, which can be really annoying. Usually choosing "Last hour" and only checking off "Cached images and files" works for me. There is no need to clear all history, logins, etc.

3. Only if it's still unresolved do you need to do a full delete of all browsing data, close your browser, reopen it, and log back in to Birst.

StefanTaganyi

I have to agree with the other CSD Analytics customers here. From a customer experience point of view, the CSD analytics prebuilt content is a letdown. The single biggest issue has been performance and most of our users don't use it because of that (some of the missing joins are also mindboggling). That is not how it's portrayed in the sales cycle unfortunately. We've heard the promises of performance improvements but have yet to see the true benefits. One change that has been good is the subject area caching flag, but you have to get infor support to enable it. Now we just need more when it comes to performance and fingers crossed that what Yasmin Barnes mentioned will in fact make a difference with the release of the June update.

I specifically mentioned prebuilt instead of Birst because I was made aware by a Birst consultant that the problem is the way the prebuilt was architected. The slowness is directly related to that and custom spaces built for what you need will be lightning quick compared to it.

Alternative Set Extract Date Hierarchy.rtf

Marie Baker

Thank you for adding that Browser expertise.

Copyright © 2025 Infor. All rights reserved.