The SSAS Usage Totals tab in SentryOne SQL Sentry provides information that can be used to determine what responses to take based on activity levels. This information is divided into three different groups:
SSAS works efficiently with default settings right out of the box.
However, when you're dealing with factors such as large databases, many concurrent users, insufficient resources on your server, or a SSAS database design that doesn't follow best practices, you can run into performance bottlenecks and problems.
As a DBA, do you have enough detailed information on the data that is in use in your SSAS Multidimensional or Tabular instance? Can you find data, such as aggregations, that your SSAS Multidimensional instance isn't using, as these may negatively impact SSAS processing and storage?
If you're struggling to troubleshoot SSAS performance problems with insufficient information, take a look at SQL Sentry. SQL Sentry gives you the visibility you need to create and maintain a high-performance SSAS environment.
For SSAS Multidimensional instances, SQL Sentry provides three tabs to help you analyze data usage.
The first tab, Attributes, consists of attribute combinations—such as date, postal code, country, and so on—in SSAS commands, along with information such as duration and the number of queries used in a particular attribute group.
You can expand each attribute group to see various queries along with pertinent information for each query, including its duration, CPU usage, Formula Engine time, Storage Engine time, and so on.
This information helps you find queries that might be good candidates for aggregations.
The Aggregations and Partitions tabs in SQL Sentry provide, similar to the Attributes tab, information about their usage. You can use these two tabs to see if some aggregations and partitions are never being used.
Because aggregations can be expensive resource-wise, SQL Sentry helps you decide if a particular one is necessary.
In SSAS Tabular instances, the SSAS Usage Totals shows you one tab, Object Memory Usage. You can view the objects that are using the most memory. As you drill down, you can see how much of that memory is dedicated to the data and the dictionary as well as see a hierarchy or the relationships to other columns.
Another piece of information you need to successfully troubleshoot SSAS performance is the cardinality of your data. Cardinality refers to the number of unique values. The SSAS VertiPaq columnar storage can realize up to 10x data compression.
However, this rate depends on your data’s cardinality. Data with low cardinality (such as gender) will see the best compression. Data with high cardinality (such as invoice number) will see the worst compression.