Skip to main content
You are viewing the documentation for Interana version 2. For documentation on the most recent version of Interana, go to


Scuba Docs

Interana 2.22 Release Notes

This applies tov2.22

Interana release 2.22 includes the following new features and resolved issues.

See Release 2.22.1 for information about the latest maintenance release. 

New features

Derived columns static string support

Derived columns can now return string values, in addition to integer and decimal types.  These strings must be static, non-mutating values (i.e. a known set of literals).

This new type can be accessed in derived columns by using static_string. The input column does not have to be a string column.

For example: 

static_string foo() {
  if (funded_at < 1199174400000) {
    return string_id("ancient");
  } else if (funded_at < 1357027200000) {
    return string_id("old");
  return string_id("recent");

See Derived column syntax reference for more information. 

Decimal support in metrics and first/last aggregators

Interana now supports decimal (float) data types in metrics and first/last aggregators. See Metrics, Measures, and Aggregators for more information about first/last aggregators.

Configurable query timeout

In previous versions of Interana, every environment had a hard-coded query timeout of 180 seconds and queries that exceeded this threshold would not complete. In 2.22, we have added a configurable query timeout parameter that can be updated on a per-cluster basis. This new parameter is called query_timeout_ms and allows you to set the maximum query completion time before issuing a timeout to the user. 

To change your limit, please contact your Customer Success Manager or send an email to

Import resiliency with data node loss

Interana can now continue to ingest data even when a data node (or multiple data nodes) becomes unavailable. Data that would normally be written to unavailable nodes will instead be written to one of the available nodes in the cluster. When the original destination node becomes available again, the data will be copied back to that node. In previous versions, ingest would stop entirely if any data nodes in the cluster were unavailable. Note that data will not be available for queries until it is written to its original destination node.

By default, this feature is enabled for the loss of one data node. There is a configuration to control how many data nodes can be unavailable before ingest stops completely. The number can be increased from the default of 1, or set to 0 to disable this feature. This feature also requires a configuration to set the destination path for data copied to temporary nodes when the original destination is unavailable.

To change this setting from its default behavior, please contact your Customer Success Manager or send an email to

String server rolloff

In this release, when an administrator chooses to roll off data prior to a certain date, Interana has the capability to also roll off all strings that were last referenced before that same date. This is subject to the restriction that Interana began tagging strings with a "most recent reference" timestamp in version 2.20, so any strings imported into the system prior to the 2.20 version will all be tagged with a timestamp based on when Interana was upgraded to version 2.20.

Deleted strings disappear from the UI immediately, but are removed from disk the next time you restart the Interana server.

Configurable auto-detection of hex columns

We added a new parameter (force_hexn_to_string). If this is set to 1, Interana will automatically detect new hex columns as strings during data ingestion. 

Regex in "text contains" filters

The "text contains" filter in the Explorer now accepts arbitrary regex. For example, the following are now valid entries:

<column_name> text contains [Aa]pple #matches both Apple and apple

<column_name> text contains ^[A-f].* #matches anything that starts with A-f

Resolved Issues

  • Custom ratios with Count Unique in numerator do not render in Table View
  • Multiple measures grouped by hex column fail in Table View
  • Cannot join lookup tables on hex columns
  • Order By is case sensitive
  • Switching from Time to Bar View sometimes causes empty chart
  • Decimal averages fail in A/B View

Known Issues

  • First/Last aggregators on strings do not work when used in Cohorts
  • Tooltips are not available for custom measures 
  • Cannot edit named expressions through links in data tooltips if the name includes an apostrophe
  • Sampled stacked bar query results are not always correctly ordered
  • Cannot share charts with email addresses that contain the plus (+) character
  • Exporting from Bar and Pie View may show JS codes in header
  • "X days before" syntax does not work for X > 12
  • Exporting from Stacked Bar View is broken
  • Pinning others' unpublished metrics to your dashboard may break the dashboard
  • Switching from Table to View View can put an unexpected ceiling on time resolution
  • Changing chart controls kills the auto time resolution behavior
  • Cannot export non-ASCII characters to CSV
  • Chart legends in Distribution and Stacked Bar View do not toggle
  • Common filters on dashboards break if a chart contains a native advanced filter

Cannot change Resolution to less than 1 week with group by in Time View

In some cases, users are unable to change the Resolution setting in Time View with a group by applied. A user will appear not to be able to get out of this state.


In the Chart Controls, set Showing Top to a value other than the maximum value and click Apply. After the query completes, you can change the Resolution setting for the visualization. 

"Too few leaves responded" error on small clusters

On small clusters with few nodes, Interana returns a "too few leaves responded" error when one data node is not available. This occurs because if a data node is unavailable and there are only a small number of data nodes, the system can't scale the results properly and will not run a sampled query.

For larger clusters with more nodes, Interana will scale values from the available nodes, and can complete a sampled query even if a data node is unavailable.

Release 2.22.1

This maintenance release fixes the following issues:

  • Data ingest process returns a "403 Forbidden" error after upgrading to 2.22
  • Incorrect results when using Count Unique with a hex column shard key
  • Queries attempting to match a large string column return an "indirect rpc call" error
  • Events with the same hex value are imported into different folders after upgrading to v2.22; this caused Count Unique queries to return incorrect results
  • Cannot convert string columns to hex without deleting the table's metadata
  • Some string values are not displayed when using the Last aggregator with Table view
  • Was this article helpful?