After 270+ Pull Requests from 34 contributors Redash v7 is here.
What’s new in v7?
Below you can see some of the highlights in this release.
Besides all the features, bug fixes and improvements listed below we managed to convert a large portion of Redash’s frontend code from Angular.js to React. v8 is going to be Angular free.
All data source options are now encrypted in the database. By default the encryption uses the
redash.settings.COOKIE_SECRET), but you can specify a different value by setting the
REDASH_SECRET_KEYenvironment variable value. Note that you need to set this before doing the upgrade.
- New Data Sources: Uptycs and Apache Drill.
- Snowplow: is now enabled by default & supports region setting.
- Elasticsearch: add support for Amazon Elasticsearch IAM authentication (with IAM profile or key/secret pair).
- PostgreSQL: add support for serializing range values.
- Redshift: remove duplicate column information for late-binding views.
- Athena: load all databases (using pagination).
- BigQuery: correctly handle temp tables with no schema field.
- Jira (JQL): support for fetching all records with pagination.
- Prometheus: fix schema loading and add support for query range.
You can now open the Knowledge Base inside the application. We also added a few “help triggers” in the app, that will open the Knowledge Base in context of what you’re currently doing.
Dashboard Parameters: We improved the flow of adding queries with parameters to dashboards and now give you full control over how parameters are mapped. You no longer have to make sure all parameters have the same name or use the
Globalcheckbox. We also added new options, like keeping the parameter local to the widget or setting a static value. Read more in our Knowledge Base →
- We added server side validation of parameter values for all parameter types, except for parameters of
texttype. All validated parameter types are considered safe. When a query is using safe parameters (or no parameters at all), View Only users can refresh it.
- Refreshing safe queries is done using the new results API endpoint, which takes only a query ID (and optionally parameter values) and does not need the query text.
Query Editor Improvements
- Run only the highlighted query text: hit Execute after highlighting a portion of your query and only the selected portion will be sent to the database. This is useful for testing sub-SELECT statements and CTE’s.
- Improved auto complete: add a dot . after a table name in the query editor and auto complete will only suggest columns on that table.
- Autosave parameter configuration changes.
- YAML syntax support (for data sources like Yandex Metrica).
Improved Query Scheduler
The Query Scheduler got a face lift and some new options: you can pick a day for a weekly schedule to run on and also set an end date after which the query will no longer execute on schedule.
We added Apache Drill, Uptycs and a new JSON data source. Also fixed a few bugs in Athena’s query runner and others.
The users page got revamped with a new look and feel and few new features:
- An indication when a user was last active.
- Show if an invited user hasn’t finished the setup process yet (Pending Invitations section).
- You can now generate a new API key for users, if there’s a concern it was compromised.
- New Celery queues status screens, replacing the old Queries Status and better reflecting the status of running queries.
- Make the queue name for schema refresh job configurable. The default used to be hard coded
schemas, which is not available on all setups. Now it’s
geventlibrary is installed by default, and you can now setup gunicorn to use
- New Docker entrypoint command to do a health check for a worker process.
- Flask-Admin is no longer setup or supported.
- New Alert destination: Google Hangouts Chat.
- When downloading results from the results API it will set a user friendly filename for the downloaded file.
- Archived Queries section added to the queries list.
- Fixed: fork query does not fork tables but instead adds default table.
- Fixed: when deleting a visualization, any widget using it was left empty on the dashboard.
- Fixed: issues with Query Editor resizing on new versions of Chrome.
- Fixed: issues with exporting dictionaries to Excel.
- Fixed: Cohort visualization gets stuck when passing string values.
- Fixed: use series name for Pie chart label.
- Make sure Flask app created in Celery’s worker process (could cause some query runners to get stuck while running queries).