Release Notes - 2021.05.12

As the next step on our strategy to bring more OneAI capability into Storyboards we have a couple of subtle changes to the Storyboard Design Panel in this release. Firstly, the panel is now called Tile Settings as the options are more than just design options and when applicable we will start showing a new tab called "Discover" with all existing options under the "Design" tab. Discover is where you will see more advanced OneAI options appearing in the future. Today the Discover tab has the existing option to enable a forecast to run by default. Ensure your seatbelts are fastened as coming releases will start to reveal more exciting features here. (ref 8337)

We have improved the Storyboard design experience to better handle multiple users modifying the same Storyboard at the same time. Today if more than one user is making changes to a Storyboard then the last changes that are made can take precedence over changes another user was working on as the whole Storyboard is saved each time. As of this release, we will only save the specific changes a user makes during the editing session. Naturally, if users are working on the same tile then there can be a conflict, but hopefully the new enhancements will go a long way to improving the experience for multiple Storyboard editors. (ref 7861)

In order to improve visibility on how up to date data is on site, we've added scheduling to all of our File data sources. These can be configured the same way as schedules for any other data source:

These schedules will automatically create a Data Load when new files are expected:

At the end of the “Time to receive files (minutes)” window, the data load will be marked as Incomplete if not all files have been received. An error message is also included to identify which files were missing:

This allows users to check retrospectively to see whether data was received at the time it was expected, and which files specifically were missing. It also remains possible to use File Data Sources without a schedule, and to send files outside of the schedule. Both of these use-cases will work as they did before the release. (ref 8124)

We've added a Last Updated On date to each file included in a Data Source so that it's clear when each table in the data source is updated. This helps to identify which tables have not been up to date in the Data Sources page. (ref 7368)

Note: This date will be automatically updated for data sources that are using our Data Pipeline V2 services, which are the services that load data from files in the data source into Redshift. In addition to the Updated Date, these new services also have improvements built into them for scalability and reliability, and are in the process of being rolled out and tested across the broad range of data sources within each of our customer environments. You can see whether they have been enabled by looking at the Use V2 pipeline services flag in the configuration for each data source:

We've found that the Oracle APIs can on occasion fail to return an accurate list of Worker IDs, and instead return some duplicate Worker IDs while missing other Worker IDs altogether. We've implemented a feature that compares the number of Worker IDs we expect from Oracle to the number of IDs that have been returned, to ensure that we now have an accurate list of Worker IDs. (ref 8263)

In order to improve stability for large data sets, we've now batched the records being returned from the SuccessFactors API. This will allow the API Run to be retried much more easily if there is any instability on either the SuccessFactors, or One Model sides. (ref 3129)

File Names for SFTP Data Sources no longer need to be case sensitive. (ref 7795)

We fixed a number if issues with the new data visualization design capability, including:

We fixed an issue where Time Projected Metrics wouldn’t work at the day level. (ref 7098)

Fixed an issue where Time Projected Metrics were not applying Time filters correctly if the filter was a different level to the Time Selection. For example, if the query returned a Time Projected Metric at the Year level, but Q2 of a specific year was filtered out, the Q2 result would not be filtered out. (ref 8314)