Qlik Replicate Scalability Options

[pedromartinsdxc]

Dear Team,

Our Replicate deployment has been very successful and stable over the last 12 months, and we are now reaching the limits of our physical infrastructure.

Regarding options to scale up the infrastructure, what is the best way to proceed:

- Scale up i.e. add more CPU/RAM to the current machine;

- Scale out i.e. add more servers to the deployment

Is there some official Qlik documentation on this topic?

Thank you,

Pedro

[Ola_Mayer]

The answer to your question is it depends 🤔

For example something to look at is how many tasks you have allocated to each server?

Are all Tasks are growing in memory/cpu consumption or is it just one?

I suggest you start analyzing you environment using Enterprise Manager Analytics and see how the memory/cpu utilization is divided among tasks sharing the server.

[pedromartinsdxc]

Thank you for the feedback! We are approaching 40 tasks in our deployment.
Most tasks have a similar level of resource consumption but a small number consumes very little resources. Is this normal?

[JitenderR · Accepted solution]

@pedromartinsdxc  To close the loop on this discussion, you also need a thorough understanding of Replicate architecture, the endpoints you are working with. For example, if you are reading a single table with 3 billion records and expect 500 million changes in a day, then a lot of performance tuning needs to be done and the task having this table itself might consume most of the replicate box resources. Best way is to engage CSE/Qlik Consulting Services for a thorough review of the environment. 

If the above answers your question, please accept this as a solution so that we can close the thread. 

Regards

JR

[JitenderR]

@pedromartinsdxc  Yes, by resources if you mean memory/CPU, if the number of incoming changes are very less, then they will consume less resources. 

Regards

JR