2 hour time limit scheduled reports - how to avoid?

[po]

Hi,

We have power bi pro and have seen gateway timeouts due to 2 hour limit being breached.

Have a few questions

1)   Any plans to increase this 2 hour time limit for pro users?

2)   The import seems to do it in steps of a few thousand - any ption to increase the amount of rows it gras in each fetch

3)  We have already opimised data model and cut down volumes in account of this but woud be goodif could optimise with view to increasing volume of data we can import

4) Direct query not option here believe as oracle version 11

5)   Does the 2 hours include the evaluation stage e.g if takes a few mins to evaluate the view?

6)  Any other optimisation tips?

Thanks

[v-deddai1-msft]

Hi @po ,

Unless you transfer your workspace from shared capacity to premium capacity, you cannot increase the refresh limit time(for premium capacity, it will be 5 hours). But  it cost extra money for premium capacity. And you can use incremental refresh, reduce the amount of refreshed data, reduce the refresh time.

In import mode, power bi will use twice as much memory space as the current dataset to refresh (the entire data set will be imported into memory), so lack of memory in the shared capacity is also one of the reasons for the slow refresh (cannot increase Memory size in shared capacity). In addition, there are some reasons: too many datasets are refreshed at the same time occupying CPU and memory, the data source system has poor responsiveness, network delay, and insufficient gateway throughput.

You can optimize your data model from the following suggestions:

Data source optimization: Appropriate use of optimized content such as indexes, reduce the addition of calculated columns and tables in the data model, display these contents directly in the data source, ensure the resources available to the gateway, increase bandwidth and be close enough to the data source.

Model-level optimization: reduce complex conversions, ensure appropriate data source privacy levels, achieve query folding, avoid loading full results, and generate cross-query combined results; reduce the size of the data model and delete unnecessary columns and rows; reduce the use of bidirectional filtering ; Use aggregation tables in large models and compound models, load pre-aggregated data to achieve fast query response; adjust DAX expressions for metric values and RLS rules; incremental refresh (Premium).

Best Regards,

Dedmon Dai

[po]

Hi,

Thanks for reply.

Regarding insufficient gateway throughput -- how can we increase this - when run report in desktop can see takes long time e.g just goes up in thousands n each fetch if we could increase this that could help.

Thanks

[v-deddai1-msft]

Hi @po ,

It can't be changed. Would you please try to follow the options i provided above.

Best Regards,

Dedmon Dai

[po]

Thanks for update

[Greg_Deckler]

Would need to see the Advanced Editor code. But, some things that tend to add no value and can take up time, reordering columns, usually sorting, things like that. There is a recent blog article about using Table.View to cut refresh times in half, might be applicable to your situation. It was on @GuyInACube just recently.