We are pleased to present our latest offering of pi - our April 2023 release! This page helps summarise the areas of the dashboard we have Added, Changed, or Fixed meaning you can make a quick and informed decision on when to upgrade to the latest release.

For our piSaaS customers, we also include any platform changes we have made to piSaaS as we continue to evolve the solution.

Here’s what’s included in the release;

Changed

We have now removed the functionality that deals with HSQL to MariaDB migration from the migration-jar project. This migration option will no longer be available to be ran in the command line.

The maximum number of columns that the Dashboard will allow to be added to the ‘Table’ chart type has been increased from 100 to 200.

Upgraded Athena JDBC Driver to version 2.0.28:1000. In addition upgraded JDBC drivers will now contain an info file alongside them to provide info.

It is now possible to host custom fonts locally in the Pi application to allow them to be accessed and used offline. All of the standard fonts supplied with the dashboard are now hosted locally as standard. When adding a new font, it is now possible to download the necessary files, save them in the Pi application and use the locally hosted files to provide fonts in the dashboard. It is also still possible to use a URL for a font API as previously for adding new fonts.

For more information on how to add and remove fonts see:[[LINK TO NEW DOCS PAGES]]

/wiki/spaces/DEV/pages/1420787713 & /wiki/spaces/DEV/pages/1420787823 - Adam Voakes to update the master documents on piDocumentation before making these release notes public.

Fixed

Fixed an issue where the secret key was not removed when exporting an Athena Data connection. Now the secret key will be displayed as a randomised string in the export. Once the user imports this connection, the data connection will be shown as invalid due to secret key no longer being valid. When updating the existing valid data connection, the secret key will be maintained.

Fixed an issue where connecting to a MariaDB database with the “Use SSL” option enabled would fail. Now, attempts to connect to a MariaDB database through SSL are configured correctly.

Fixed an issue with the horizontal scrollbar not working on iOS devices due to the changes in one of the iOS upgrades which removed the support for the custom scrollbars. This has now been fixed and the horizontal scrollbar should become available when the user touches and drags under the table.

The changes affects the standard, rotated, editable and crosstab tables that have measures. Please refer to the ‘Known Issues’ section below for the limitations.

Previously, when drilling from a chart to a category that has the same dimension object as a drop list picker the value was not being inherited so no filtering would take place and the drop list would instead act as if no/all options had been selected. Now, drilling to a category will correctly apply the relevant filter to the drop list and all applicable charts in the category will be filtered by the expected value.

Handling of special characters in certain tooltips originating from i18n language files has been improved.

Previously: pie, doughnut, variable pie and variable doughnut charts would only conditionally re-size to take up the full space of their container - while the container is shrunk to a smaller size - if the container was shrunk BOTH horizontally and vertically. It now correctly resizes when the container is shrunk EITHER horizontally of vertically. See below how the container is shrunk vertically, but not horizontally, yet the chart takes up the space of its container correctly.

Previously: when merging tables where all dimensions are shared between all the tables being merged, it would only merge on the first matching dimension and leave a bunch of empty cells in the table result. See below:

We have now added a new algorithm which - upon the condition of all of the dimensions shared between the tables being merged have the same name and are coming from the same data connection - will now merge all of the matching dimensions together correctly. Additionally, it does not duplicate any measures shared between the tables being merged as it needs to when all dimensions do not match. See below:

This continues to work when you have filters/sorting to change the order of the dimensions, and if you have the dimensions/columns in different orders/positions on different tables. It works with any number of dimensions so long as they all match. It works with 2 or 3 tables being merged.

On top of this, when using this new algorithm, we have fixed a bug where if an additional data column exists in one table but is visible in another being merged, it correctly includes the column in the table - see the final column of the merged result - this bug, however, still persists when using the old algorithm where all dimensions do not match.

Known Issues

If you have a crosstab table that contains dimensions only, the horizontal bar will be invisible and some content could be missing if the chart container is too small to fit all the columns.

When using a mobile device to view the content in the dashboard (either iOS or Android), the horizontal scrollbar will not be available when viewing merged tables.

piSaaS Platform

26 Apr 2023 Awaiting confirmation from Blake that these are finalised - taken from /wiki/spaces/DEV/pages/1437040644

As DevOps, I have a safe way to terminate a pisaas instance production environment

When a Pisaas instance has been stopped there is now a way for administrators to completely terminate an instance. This will result in a loss of data but backups will be kept for 30 days. It will remove any resources the instance has been using in our production environment.

As a sysops, i have one load balanced renderer which is used to support all pisaas instances

Instead of each Pisaas instance deploying and using its own renderer, there is now one renderer for all instances. This is a larger renderer which should be able to cope better than an individual one, improving overall rendering performance.

As a trypi/pisaas user, I have one pirana endpoint to handle all requests

Like the renderer, there is now one pirana to be used by all pisaas instances to improve performance.

As a customer, i can bring my own repository database

Introduced functionality that enables dashboard repositories to be uploaded and restored from. The process includes validation checks to make sure the repository can be restored. This is administrator functionality, therefore customers can request a repository to be restored by sending this over to our Service Desk where it will then be uploaded onto the appropriate instance.

As a developer, I can request a pisaas duplicate environment based on an existing running service

Added functionality to allow duplication of an existing Pisaas instance - to be used for debug purposes. It will create an instance under a new service name with the exact configuration and repository used by the original. It will be using a new database but with the restored repository, and the username and password of the dashboard will be set to default. Permission from a customer is required before duplicating an instance for data security purposes.

As devops, i can see how "available" a pisaas instance has been over a given period

Additional reporting capabilities on Pisaas instances available to Panintelligence internal dashboards. This allow Panintelligence to view the percentage of availability of each instance, highlighting any instances which may be a cause of concern. The data will not be real-time as it is pulled back once a day, but the data received is by the minute to accurately catch any instances we may have temporarily lost.

As a pisaas administrator, i want the efs storage to be mounted to the ec2 bastion automatically on log in

When connecting to EFS on the Panintelligence ec2 instance, the efs storage is mounted automatically. This automates the connection process to make accessing instance storage more efficient. Alerting has been enabled should there be any issues during the mounting stage.

As a new user, i have a document that details the first steps into pisaas

- Coming Soon