Interana release 2.21 includes the following new features and resolved issues.
See Release 18.104.22.168 and Release 2.21.2 for information about maintenance releases.
Use the new global filters to build a set of filters that you can apply to every query. You must have the Publisher or Admin role to create and edit global filters. However, if a Publisher or Admin has defined Global Filters, any user can see and add or remove the filters in the Query Builder. (If no Global Filters have been defined, the Global Filters field will not appear in the Query Builder.)
You can also set them as default, which means they will be applied automatically to any queries your users run (global filters are defined and applied per dataset). As with regular filters, you can use the basic or advanced syntax to define the filter parameters.
To create or edit global filters on your Interana instance, go to
See Create and use Global Filters for more information.
If the filters applied in a Global Filter conflict with those in the Explorer, arguments will be treated as AND (and therefore return no results).
First and Last aggregators
We added First and Last aggregators that return the value of a given column with the smallest or largest event timestamp.
You can use the First and Last aggregators when creating named expressions. When working with int columns, the aggregators are included in the Measure list for the following views:
- Stacked Area Time
For string columns, you can use the aggregators as measures with the Table view.
See Metrics, Measures, and Aggregators for more information about the first/last aggregators.
Advanced filters in named expression builders
We've added an advanced filter option to the named expression builders. You can filter using advanced syntax, including OR statements, when building cohorts, metrics, sessions, funnels, and global filters.
Your advanced filter will be partially validated for correctness when you save a named expression. We will notify you about syntax errors immediately, but referring to an invalid column name will not show up until you attempt to run the named expression.
Group By for sets
You can now group by integer and string sets when building queries. Grouping by a set column will aggregate individual elements of sets across events in the query. Sets are assumed to be unordered and can be comprised of integers and strings.
Because event data can exist in multiple sets, the sum of the events of the set elements will be greater than the actual number of events.
See Group By Sets examples for more information.
We added tooltips to column names and named expressions in lists. The tooltip pop-ups include more information about the data, based on the Description field of the columns and named expressions. This will help you select the right data when building your queries.
You can also click Edit in the tooltip to open the column detail dialog and add or edit a description.
We added an Embed Chart option that lets you use HTML to embed a dashboard chart in a web page or tool that supports HTML. In the Explorer window, select More > Embed Chart, then copy the HTML code to a web page.
Improved display of columns in Table view
We have improved the display of information in Table view. In previous releases, if you selected multiple Compare Groups, we included that information in a single column. Now we display separate columns for each group.
New Order By chart controls
We added an Order By chart control that lets you set how data is displayed in Bar, Pie, and Stacked Bar views. You can select to order by the measure or by any groups that you selected, and in either ascending or descending order.
When Ordering By any Measure in Time View, the group labels below the chart are sorted by the first value in the time range.
Editing restrictions for datasets
Only users with Publisher or Admin roles can edit dataset properties, including column descriptions and familiar names.
Decimal support for derived columns
You can now create derived columns that return decimal (double) values.
"Last x" support for time ranges
We support both precise ("now", "1 day ago") and calendar-aligned relative time ("today", "yesterday", "last week"). See Time query syntax for more information.
Support for the Apache Avro file format
We added a new ingest transformer (avro_load) to support ingestion of files in the Apache Avro format (https://avro.apache.org/).
Interana Install creates the "customer" by default
In prior releases, after installing Interana it was necessary to run the "create_customer.py" script as a final part of the setup process. Starting in release 2.21, an "interana" customer is pre-created during the install, and instead there are a few distinct steps needed to bootstrap the initial admin user and configure auto-registration rules.
# provision the initial admin user /opt/interana/backend/deploy/user_admin.py signup -c1 -u email@example.com -p test --base_url "http://localhost" /opt/interana/backend/scripts/rbac/user_role.py create -c1 -r admin -u firstname.lastname@example.org # enable e-mail auto-registration if desired /opt/interana/backend/deploy/update_customer.py -c interana --add_email_suffix interana.com
In previous releases, the uncompressed tarball for upgrading could be located in the home folder (~/). But in version 2.21, the uncompressed tarball must be in the
/opt/interana/backend/ folder of the push node.
- Can't explore off a dashboard chart that reads "current not available"
- Using multiple "text contains" arguments in a cohort will return an error
- Typeaheads for cohorts don't work when Publishing features are enabled
- Average aggregation inside metric returns integer instead of float
- Decimal columns show as data type "int" in the Dataset Settings page
- The "text contains" filter does not accept regexes
- Scrolling is broken on dashboards when scrolling over Table charts with scroll bar
- Sorting by "Last Modified Date" in named expressions uses string instead of date sorting
- delete_columns can take 1 minute to execute for single shard
- Tooltips are not available for custom measures
- Changing the familiar name of column doesn't show up in type-ahead until refresh
- Order By is case-sensitive for the Stacked Area Time view. Items that begin with capital letters are sorted before items with lowercase letters (Interana sorts A-Z, then a-z).
- Viewing sets in Samples View does not work for sets with > 10 values
- Unable to use count unique with set columns
- Decimal average fails with error in A/B view
- Cannot run unsampled queries in Chrome v43
- We calculate 6 digits of precision but cut off trailing zeroes
- Cannot change the resolution when you visit open a query from a dashboard if that query is using a Group By
- Cannot filter to or filter out a decimal value if the query groups by decimal values
- Order By does not work with custom metrics
- Cannot edit named expressions if the name includes an apostrophe
- Cannot use derived columns in global filters
- When adding multiple filters that use the same column and filter syntax to global filters, they are combined with the AND syntax instead of OR
- Firefox only: cannot scroll large tables. The workaround is to hover your mouse over where the scroll bar should be on the right side of the window, start to scroll until the bar appears, then click and drag the scroll bar to scroll through the table.
- The chart legend cannot be toggled when using the Distribution or Stacked Bar views
This maintenance release fixes the following issues:
Dashboard charts with stacked bar graphs appear as empty charts in email reports
Switching from Basic to Compare filters in the Explore window does not save the filters to compare group A or B
Order By does not correctly sort Custom Metrics in the display
Queries using a custom ratio with "Count Unique <string column>" in the numerator return no results in Table view
Can't use Make Fraction of Total with Count Unique in the numerator when creating a ratio metric
Removing a "true" or "false" filter values does not update the query correctly. New queries still use the true/false value.
This maintenance release fixes an issue that could occur when tiering data.