Yellowfin 8.0.2 is now available and with it brings an assortment of enhancements and fixes. A major area of focus in this release (and for the rest of 2019) is overall performance within the analytics suite.
We have introduced a number of improvements to loading times in the following areas:
- Browse Page
- View Builder
- LDAP Authentication
However, one thing to note is that performance improvements will not always be apparent to everyone. Many things impact performance, including the underlying infrastructure that Yellowfin is installed on, the performance of the reporting databases, network latency, the amount of content an individual has access to, and so on. Some users may notice significant improvements (with speed increases of up to 4x) whilst others may not notice a big improvement.
Let’s have a look at what has improved in 8.0.2:
Dashboard loading times have been reduced by optimizing calls to the Yellowfin Repository database. This generally involves bulk loading metadata that was previously loaded in individual requests, and caching system information that is accessed often. These bulk loading and caching optimizations have been implemented in the metadata loading process for both Dashboards and Reports, specifically:
- Group membership, content records, and content favorites for users
- Report Canvas objects and Alert-based widgets
- Storing and caching a Dashboard’s previous IDs
- Dashboard configuration
- Discussion elements
We’ve improved the loading times for Timeline by merging requests to the Yellowfin Repository database and improving the user access determination process. By also optimizing the Signal loading process, a significant proportion of processing time has been saved on the Timeline.
Image caching at the browser level has been enhanced with Content E-tags being used to determine whether an image has changed recently and if a new version needs to be fetched from the server.
Improvements in this area also decreased loading times for notifications on other content like Signals.
By improving the user access determination processes, we have also decreased loading times for the Browse Page. Caching user group membership within this process also reduced requests to the Yellowfin Repository database.
Signal loading speed has been dramatically improved by caching the Column Formatters used for displaying data and visualizations in generated Signals. This affects and decreased loading times for both Signal notifications on the Timeline and browsing for Signals on the Signals List page.
We have introduced new configuration options for bulk loading table and column metadata from your reporting data sources. The source database metadata can also be cached indefinitely if necessary. These additional options allow for the View Builder to run much faster in certain scenarios, especially when dealing with:
- Very large views (with many table elements including joins)
- Databases that have many schemas
- High internet latency between Yellowfin and your reporting data source
If your reporting scenario is contained within these points, please contact our support team to know more about these configuration options.
We have improved the Client Org determination process for LDAP authentication. This provides decreased login times for Yellowfin deployments which are both multi-tenant and utilizing LDAP for authentication.
Group Flattening, which occurs when an LDAP user logs in (to determine access for the current session), is now calculated based on the user, rather than flattening each group the user is a member of. Groups are now patched to reflect the appropriate access for an LDAP user, if required, which improves login performance considerably.
Performance on sharing content was previously affected in Yellowfin instances with a great amount of named users. With bulk loading and optimized access processing, we’ve increased the performance on loading and accessing the share popup.