Alteryx becomes terribly slow when doing any small operations

[Kurohits]

Hi Experts, 

I am using Alteryx within citrix environment. Any updates or changes or small operations are taking hours. It is just keep buffering on screen. The database is lying on hadoop environment and when I am trying to update connection strings using input tools or even trying to update sql queries, its taking hours to execute the opertions. 

What could be possible reasons of being alteryx working slow? Is it database or RAM issue? 

How I could make it work faster? 

Appreciate your help in advance. 

[KGT]

By default every time you click on something, Alteryx will refresh the metadata. It shouldn't cause delays like your saying, but if you have 16k columns for instance, that's going to hang. 

In "User Settings > Advanced (tab) > Disable Auto Configure" will stop that metadata refresh and you will need to either press the tick at the bottom left of the configuration panel or hit F5 whenever you want to refresh the metadata.

The other possible problem, and not sure if it happens anymore, but Citrix used to used shared resources that would end up allocating about 1GB RAM to the box. That's useless for any real operation that involves further memory "hogging" and so that could be the issue as that allocation is not planning on a piece of software like Alteryx trying to allocate it's own memory management.

Lastly, the other thing I would try but should only affect Runs, is toggling AMP under runtime settings. AMP allows parallel processing and there's a chance that you could be running into thread collision or something on shared cores. I give this one a low priority.

[Kurohits]

@KGT 

Thanks for your response. 

I looked at the data and it has 259 columns only. The data is coming into Alteryx through odbc connection using hadoop environment by writing SQL queries.

I tried disabling auto configure however it is still working exteremly slow. Regarding 2nd option, how would I check and confirm if citrix memory allocation? There is a seperate team who manages all infra. 

3rd option I am not using AMP engine actually. 

I want to go into the depth of this issue and resolve it if possible. Can you please guide me how to investigate and resolve  

[Deano478]

@Kurohits you should try enabling Data Profiling in alteryx to see what tool or tools this issue is stemming from. To do this go to: User Settings > and check the box next to Collect and display data profile information. Alternatively the issue could be with Citrix itself. 

[KGT]

Yep, performance profiling will help isolate the issue. On the Runtime tab, tick "Enable Performance Profiling". The output log will tell you how long each tool takes (It's not all segregated though).

The first thing I would try is to see whether it's the input that is slow. An Input tool and then a basic operation like Left(Field,10) followed by output/browse will tell you whether the issue is solely connection or a mix of things.

[Kurohits]

@KGT 

Thanks for your reply again. 

After opening the workflow in canvas, I try to change the connection string to point to hive database in development environment. The moment I tried to set up a connection, select hive odbc and then select saved odbc connection then afterwards alteryx keeps buffering endless. I have enabled data profofiling option however this option is worthless until I run the workflow which I can't do because of connection string is taking hours to update. Also when I am doing any basic operation locally like join, left etc., its working smoothly.

Is there any way to fix this issue permanently?

[KGT]

Oh, then it definitely is the connection. That might be worth a support ticket (my.alteryx.com) where they can troubleshoot while knowing more about what happens behind the scenes. It may be worth trying the Simba connections if that's still a thing for Hive, support will be able to guide on that though

[Kurohits]

@KGT 

Yes I also think same. To raise support ticket I need license of my own which I don't have. I am using Alteryx on client environment and can't have access to raise the ticket. I have also just checked that Hortonworks hive odbc driver is used to make the connections but they have simba hive odbc driver also installed in Alteryx. 

I don't know the exact technical benefits or differences between these 2 drivers. Do you think that the drivers could possibly be reasons behind this slowness? 

Would appreciate your help here to fix this issue. 

[adamzampa]

Yes i also face same problem, Is anyone know about this please share with me.

[KGT]

Have a look at this: https://community.alteryx.com/t5/Alteryx-Designer-Desktop-Discussions/Advantage-of-using-the-Simba-D...

A driver for the "more standard" processes like an SQL connection doesn't do too much, but basically tells either side what the format means for the other side to understand it. When it comes to Hive, that job is huge as you're not talking about translating a table to a table. 

So, yes it could make a difference, but will it, I don't know. No-one can really say without knowing your environment. If you can narrow it down to the connection, then you need the applicable parties to investigate that connection. Namely, your IT and someone that knows about the Hive connectors.