Qlik Cache Management, Granularity, performance tuning and Long running query

[Ranjanac]

Hi All, 

I wanted to understand the below scenarios:

1. How can we recognize that there is an issue with Analytics page/Sql/load script . What will the the first step ?

2. How can we check the performance issue with the report? for example: the reports are taking long time to run or it gives an error: "Run Time Exceeds" . What to do in that case ? how can we check that ? and what is the solution ?? 

3. How to manage the cache in Qliksense/Qlikview ?

4. How we can define the Granularity in  Qliksense/Qlikview ?

5. How to do the performance tuning in  Qliksense/Qlikview ?

could you please assist with answers.

Thanks .

[Oleg_Troyansky · Accepted solution]

Hi there,

These are all good questions, perhaps a bit too wide for a forum message, however let me offer some thoughts:

1. How can we recognize that there is an issue with Analytics page/Sql/load script . What will the the first step ?

I'm guessing you recognize that there is an issue when things are working slower than expected. If it's not broken, it doesn't need to be fixed. When you do, the first step is to analyze the app for performance.

2. How can we check the performance issue with the report? for example: the reports are taking long time to run or it gives an error: "Run Time Exceeds" . What to do in that case ? how can we check that ? and what is the solution ?? 

The best way to analyze performance is to use the performance tuning tool called QSDA Pro.  It's a commercial product with a limited free option. It's the best (and almost the only) performance tuning tool in the market, and it can do wonders in well trained hands.

3. How to manage the cache in Qliksense/Qlikview ?

Cache is managed automatically in Qlik. You can set Lower and Upper limits to determine how much of server memory should be used for caching. Other than that, you can impact performance by following best practices - for example, avoiding multiple versions of the same expressions, like SUM(Sales) vs. sum(Sales). Or by avoiding complex calculated dimensions etc. It takes a bit of time to learn all the tricks. Again, QSDA Pro offers some help with many of these issues.

4. How we can define the Granularity in  Qliksense/Qlikview ?

I don't understand the question. The granularity of the data is driven by the data load script. As a developer, you decide what level of detail is necessary, and you load data accordingly. There is no special setting that defines Granularity.

5. How to do the performance tuning in  Qliksense/Qlikview ?

The best advice I can give, after recommending QSDA Pro, is to find your way to the Masters Summit for Qlik  (coming up soon, in Dublin and in Orlando) and to listen to my session on Performance Tuning. In my half-day session, I describe the most important best practices of performance and I show how to use QSDA Pro for performance tuning. My students come out of my session with a completely new understanding of Qlik performance. The most common feedback I get is "I'm going back to completely redesign all of my apps".

Cheers,

[Ranjanac]

i actually don't understand the concept of Granularity in  Qliksense/Qlikview. why we're using it ??? what is the actual use ? if you could brief me about it.

coz, i read it somewhere " Granularity is usually mentioned in the context of dimensional data structures" 

[Oleg_Troyansky · Accepted solution]

I don't think Granularity is a specific concept in Qlik. The way I understand it is the desired level of detail for your analysis. For example, if you have detailed transactions that happen throughout a day, you may analyze this data at different levels of granularity - either hourly, or daily, or weekly, or monthly. The more detailed is your data, the more resources (time, memory, hard drive space ...) it will require. That's my understanding of this concept, and in Qlik it just means that you should use the optimal level of detail - just enough for the business needs, but not too granular, to save on resources.

Cheers,