Changelog

The event picker now marks the events that matter most in your workspace, with a tooltip explaining the indicator. Switching between funnel and journey metrics keeps your conversion window, and several chart, date picker and Mitzu bot issues are fixed.

Added​

• Event relevance indicator. Events in the event dropdown carry a small marker that shows how relevant each one is, with a tooltip that explains it.
• Switch organization from the not-found page. When a link points to another organization, the not-found page offers a button to switch to it.

Changed​

• Switching a metric between funnel and journey keeps its conversion window.
• The show events button is hidden when a formula is in use.
• Event sampling during indexing is faster and more representative.

Fixed​

• Showing and hiding annotations works for every chart type.
• The date picker applies the selected range correctly.
• Journey settings are no longer lost when changing the metric type.
• New metrics start with the correct default settings.
• The homepage link points to the right place.
• The Mitzu bot no longer sends duplicate responses.

The share-with-annotations option is now a simpler show annotations switch, and two visual glitches are fixed.

Changed​

• The share with annotations option is replaced by a show annotations switch.

Fixed​

• Chips in menus no longer flicker or shift.
• Annotations display without visual glitches.

Databricks workspaces can now use array fields nested inside struct columns. Global annotations and heatmap charts, including their annotations, work correctly again.

Added​

• Arrays in structs on Databricks. Array fields nested inside struct columns are discovered and usable in filters and breakdowns.

Fixed​

• Global annotations display correctly on charts.
• Heatmap charts render correctly, including their annotations.

A small release that fixes heatmap tooltips.

Fixed​

• Heatmap tooltips show the correct values.

Event discovery in the indexing settings now works through events in batches. Embedded shared dashboards show a clear message when they are no longer available, and a set of sign-in and chart issues are fixed.

Changed​

• Event discovery in the indexing settings processes events in batches and shows its progress.
• Embedded shared dashboards show an error message when the dashboard is no longer available.
• Access-denied pages explain why you cannot open a page.

Fixed​

• Users of a removed organization are no longer caught in a redirect loop.
• Metrics no longer fail with a date timezone error.
• Heatmap charts display correctly.
• Fixed a script error that could break parts of the page.

Databricks connections now ask for a catalog up front, queries only read the columns they need, and buttons are disabled while a background task is running so actions cannot be triggered twice.

Changed​

• Databricks connections require a catalog.
• Queries read only the columns they need, which makes them faster on wide tables.
• Buttons are disabled while a background task runs.

You can create a cohort directly from a result list, and long result lists are trimmed so they load quickly.

Added​

• Create cohort button. Save the users in a result list as a cohort with one click.

Changed​

• Long result lists are trimmed so they load faster.

The dimension search page no longer loads event properties, so it opens faster.

Changed​

• The dimension search page loads faster.

A small release with timezone fixes and performance improvements.

Changed​

• Several pages load faster.

Fixed​

• Dates and times respect the configured timezone.

Charts now show a badge when their result comes from the cache, and several interface issues are fixed.

Added​

• Cached result badge. Insights show a badge when the result was served from the cache.

Fixed​

• Clicking a row in the lookup table opens the right record.
• Fixed several small interface issues.

Properties are now discovered on demand as you search for them, so large workspaces no longer need to index everything up front. The lookup page gains a run button, reindexing and a create cohort button, and the data catalog flags event tables that are out of sync.

Added​

• On-demand property discovery. Event and dimension properties are discovered as you search for them, instead of all at once.
• Search-based indexing. Indexing finds fields through search, which speeds up discovery, especially with dimension tables.
• Lookup page actions. Run the lookup, reindex its data and create a cohort from the results directly on the lookup page.
• Out-of-sync event tables. The data catalog shows which event tables are out of sync.

Changed​

• Dimension property values are searched dynamically as you type.
• Changing the property of a filter resets its value.
• The lookup page remembers the columns you selected.
• The cohort page has an updated layout.
• Discovery and several pages are faster.

Fixed​

• Filter values are cast to the right type.
• Cohorts built from dimension properties return the correct users.
• Annotations no longer fail when comparing dates with and without timezones.
• Fixed several layout and animation issues.

You can now cancel running indexing queries from the data catalog pages, and focus mode can show the events behind a chart. BigQuery workspaces also get warehouse query information.

Added​

• Cancel indexing queries. Stop running indexing queries from the data catalog pages.
• Show events in focus mode. Focus mode can list the events behind a chart.
• BigQuery query information. Running query information is available for BigQuery connections.

Changed​

• Dashboards without preview images show an empty grid, and cohorts and annotations no longer show a preview.
• The dimension page is cleaner and easier to scan.

Fixed​

• Cancelling a running query stops it reliably.
• Fixed several issues on the lookup page.

Workspaces can now filter case-insensitively, and ClickHouse connections can use HTTP. Drilling down into funnels with a breakdown on a later step works again.

Added​

• Case-insensitive filtering. A new insight setting matches filter values regardless of letter case.
• ClickHouse over HTTP. Connect to ClickHouse using the HTTP interface.

Fixed​

• Drill-down and see trends work in funnels whose breakdown is on a step other than the first.
• The date picker applies the selected range correctly.

Dashboards get a sensible default auto-refresh setting, and property value checks ignore duplicate values.

Changed​

• Dashboards use a default auto-refresh interval.

Fixed​

• Property value counts no longer include duplicates.

Dashboards and insights now show a preview image when browsing. BigQuery map field filters and drill-down breakdowns are fixed.

Added​

• Preview images. Dashboards and insights show a preview image.

Fixed​

• Filters on non-string map fields work on BigQuery.
• Sub-segment breakdowns are kept when drilling down into a metric.
• Fixed several small styling issues.

Array fields nested inside struct columns are now supported, and warehouse indexing queries can be cancelled. Indexing a single dimension table no longer removes the fields of other tables.

Added​

• Arrays in structs. Array fields nested inside struct columns are discovered and usable.
• Cancellable warehouse indexing. Indexing queries running in your warehouse can be cancelled.

Fixed​

• Indexing a single dimension table keeps the fields of other dimension tables.
• Fixed a UI glitch and an issue that stopped some actions from responding.

You can now run a query with Cmd+Enter, and Y-axis range settings are easier to use. Formula comparison charts and query information on Snowflake are fixed.

Added​

• Run with Cmd+Enter. Press Cmd+Enter to run the current query.

Changed​

• Y-axis range settings are easier to configure.
• Dimension tables are easier to set up.
• Query information on Snowflake loads faster.

Fixed​

• Formula comparison charts show the correct values.
• Right-to-left text displays correctly.
• Warehouse query information loads reliably.

You can now choose fields to ignore in dimension tables, and comparison charts show the original date of each point. The separate discovery-with-filter-values button is gone.

Added​

• Ignored dimension table fields. Choose fields in a dimension table that indexing should skip.

Changed​

• Comparison charts show the original date of the compared values.
• The discovery with filter values button is removed; discovery handles values automatically.
• The query manager shows the type of each query.
• The empty more menu is hidden on the catalog tab.

Fixed​

• Ignored columns are skipped correctly during indexing.

A patch release that fixes Databricks connections and discovery of map columns.

Fixed​

• Databricks connections work again.
• Map columns are discovered correctly.

Discovery now caps the number of string values it stores per property, and map columns are discovered correctly.

Changed​

• Discovery limits how many string values it collects per property and skips malformed values.

Fixed​

• Map columns are discovered correctly.