asked on December 10, 2021
Hi FM Team,
I am a licensed user .Kindly provide a resolution of the same .
Problem Statement :
We have Hosted the FM Data server in the Azure Environment.
However we have observed that the Hosted version is extremely Slow, even for a basic select query
SELECT * FROM [xyx].[Dim_Calendar]; Table has 5000 records only. (accessing the Azure SQL Table)
However the Local Version of the same FM Data server works faster .
Kindly suggest what is the basic memory , CPU or any other recommendations that is recommended for Hosting the FM data server .
Regds,
Aparna
⋅ December 13, 2021
Hello, Aparna,
Thank you for reaching out to us.
Our team suggests the following recommendations for achieving better performance when working with Flexmonster Accelerator:
As you have noticed, the local version performs better than the hosted one. Please consider that Flexmonster Accelerator uses the ADOMD.NET library for connecting to Analysis Services servers and querying data. The ADOMD.NET library is known for performing slower when working with remote Analysis Services servers. Therefore, placing the Accelerator on the same machine as your Azure Analysis Services instance should minimize the time spent on data transfer and, in its turn, improve the performance.
We hope this helps.
Please feel free to reach out if further questions arise.
Kind regards,
Vera
⋅ December 14, 2021
Helo Vera,
Looks like in your response you are mainly referring to the FM accelerator performance.
My query is specifically on the FM DATA SERVER extremely poor performance .
We have Hosted the FM Data server in the Azure Environment- App services .
However we have observed that the Hosted version is extremely Slow, even for a basic select query
SELECT * FROM [xyx].[Dim_Calendar]; DIM_Calendar Table has 5000 records only. (accessing the Azure SQL Table)
However the Local Version of the same FM DATA SERVER works faster with the same set of queries.
Kindly provide s the basic memory , CPU or any other recommendations that is recommended for Hosting the FM data server.
Regds,
Aparna
⋅ December 14, 2021
Hello, Aparna,
Thank you for your reply and my apologies for the mix-up.
We would like to mention that Flexmonster Data Server should be able to handle more than 5000 records just fine: the maximum data size depends on your RAM capacity and the number of unique members in the dataset. For an index containing 5000 records, 8 GB of RAM should be more than enough for Flexmonster Data Server to perform well.
It seems there might be an inbound/outbound limit imposed by the Azure Portal, which causes the slowness.
Flexmonster Data Server logs how much time was spent on processing each request (when the LoggerMinLevel configuration option is set as "Info"). Our team suggests checking the response time details in the Flexmonster Data Server logs. This will help to understand whether the slowness is caused on Flexmonster Data Server's end.
Please let us know if this helps. We are looking forward to your response.
Kind regards,
Vera
⋅ December 15, 2021
Helo Vera
5000 records is the Minimum /Smallest Index which we are trying ,which was slow .
In the Actual scenario we would require around 12 Indexes to be configured each index can have upto 10,000 records.
What is the Azure Infra recommendation from your end for the same a faster retrieval?
Regds,
Aparna
⋅ December 15, 2021
Hello, Aparna,
Thank you for your reply.
For the specified amount of records in your data indexes, 8 GB of RAM should be enough. However, we suggest running some tests on your side and checking the performance based on Flexmonster Data Server's response time logs.
As for configuring Azure infrastructure, we suggest checking the Azure Portal settings and documentation for lowering possible limits to achieve a faster retrieval. Also, from our experience, selecting a server in the closest region could help save time on the data transfer.
We hope this helps.
Kind regards,
Vera
