This document introduces new features included in Interana release 2.25 and lists newly resolved and known issues.
See Release 2.25.1 and 2.25.2 for information on the latest maintenance releases.
This section provides a high-level description of the new features included in release 2.25:
- Selective data deletion
- Privacy purge
- Parquet ingest support
- Automated whale handling
- Time handling enhancements
- String tier optimization
Selective data deletion
This feature allows you to selectively delete events within a particular time range if their column values match a list of values to be removed. There are a number of cases for selectively deleting events:
- When there are garbage records that can be identified with an "is one of" filter expression.
- When you need to delete records for a particular set of actors.
- When you need to delete records from a particular data source.
- When there is a long retention period for specific high-value events, and a short retention period for all other events. You can use selective delete to periodically delete events of a particular type that are older than a specified date.
You can preview the results of a selective delete job without removing any events and review a list of deleted events after the job is run. For more information, see Selective data deletion.
Interana Privacy Purge enables you to comply with GDPR and other privacy regulations and policies to which your company adheres. Privacy Purge enables you to protect the privacy of Interana users, as well as the users of services whose data resides in Interana.
A privacy purge enables you to delete all the events for a specified actor across the entire cluster, as well as deleting strings associated with the actor. Likewise, references to the specified actor are deleted from filters, named expressions, dashboards, and any other metadata in the ConfigDB. To comply with legal requirements, you may also need to delete all logs of activity for users who request a privacy purge.
For more information, see the following articles:
Parquet ingest support
Automated whale handling
Data whales create unbalanced data that can cause problems in a data tier. The automated whale handling feature provides a solution for detecting and removing data whales. For more information on automated whale handling, see Balancing data for efficient sampling.
Time handling enhancements
This release introduced ways to use the Interana CLI to change the query time zone and dashboard timezone offsets to account for daylight savings time. For more information, see the following articles:
String tier optimization
Enhancements were made to significantly improve string tier performance and backup recovery time.
Specify "order by" for API queries
You can now specify the order in which results are returned when querying via the API.
The following issues were resolved in Release 2.25.
|HIG-7728||Error messages for string server synchronization did not clearly state what had failed and the origin of the failure.|
|HIG-11801||Deleting a global filter broke pinned queries that used that filter.|
|HIG-12112||Support for Google authentication was not available.|
|HIG-12446||Query API structured logs rapidly consumed too much space, which resulted in decreased performance.|
|HIG-12463||The "months" text label was for 30-day month value, which caused confusion. This label has been changed to "30-day months".|
|HIG-12465||Using a funnel metric in query builder resulted in an error.|
|HIG-12488||String servers took too long to recover after an upgrade.|
|HIG-12739||Sampled count events sometimes returned a decimal value. This issue has been resolved so that values are rounded to the nearest whole number.|
|HIG-13070||Large numbers that were identifiers stored in integer columns were rounded by filters before being displayed in the UI.|
The following are known issues in Release 2.25.
|HIG-8483||After authentication, a user has to click a link twice to view the desired page.|
During a privacy purge, if two or more tables have columns with the same name that are different column types, an error results showing each table, column name, and type.
Workaround: Change the friendly_name of one or more of the columns shown in the error so that the names are unique.
|HIG-12905||Values that have spaces in the name are not supported by Interana privacy purge.|
|HIG-12917||When there is a named expression and a cohort or session has the same name as a value, filter results only show the cohort or session, not the value.|
|HIG-13086||Selective data deletion and privacy purge do not currently support decimal values.|
|HIG-13090||User IDs represented as integers that exceed 53 bits are stored correctly upon ingest, but when displayed in the UI are rounded with the last few digits shown as zeros.|
Deleted ratio metrics reappear in the UI.
Workaround: Refresh the page and the deleted metrics disappear.
The dashboard cache is not included in a privacy purge. However, the cache refreshes every week, clearing out old data. For this reason, there may be a short time when dashboards still display privacy information that has been purged.
Workaround: Wait for the dashboard cache to refresh.
Columns that do not have a specified type cause a significant impact on performance during import.
Workaround: Stop the import, specify a type for any columns that do not have a type specified, then restart the import.
This section introduces feature enhancements included in Interana release 2.25.1, as well as newly resolved and known issues.
Release 2.25.1 includes the following enhancements.
- To allow for halting a privacy purge, the
ia purge interruptoption was added. For more information, see How to perform a privacy purge and the Interana CLI reference.
- To prevent moving unbalanced shards, such as whale shards, you can now use the
ia tier rebalance --freeze-shard-idoption. You can specify multiple shards in a comma-separated list (with no spaces in between), as long as each shard is on a different node. Use this option on the data tier ONLY. For more information, see Resize a cluster and the Interana CLI reference.
- For enhanced clarity, the
ia tier rebalance --excludeoption was renamed to
--hosts-to-clear.To rebalance data to a dedicated data node, use the
--hosts-to-clearoption. You can specify multiple nodes to be cleared using a comma-separated list with no spaces in between. For more information, see Resize a cluster and the Interana CLI reference.
Release 2.25.1 includes the following newly resolved issues.
|HIG-12992||A delete pass for a privacy purge was taking too long.|
|HIG-13086||Selective data delete and privacy purge did not support decimal values.|
|HIG-13090||User IDs represented as integers that exceeded 53 bits were stored correctly on ingest, but displayed as rounded numbers in the UI with the last few digits as zeros.|
|HIG-13105||Large numbers were rounded when displayed in Table View. This is enabled on a per-cluster basis. Contact your CSM if you encounter this issue.|
|HIG-13196||You were unable to delete a table with shard function exceptions (whales) without first deleting the shard function rows.|
|HIG-13398||Selective delete failed when run on tiered storage.|
The global filter URL required an upper case F which caused confusion, this requirement has been removed. Users can now access the global filters with either of the following URLs:
|HIG-13577||In some cases, shards were not distributed evenly across data nodes during a cluster rebalance.|
|HIG-13624||In certain situations, large numbers were rounded and displayed incorrectly in the UI.|
Release 2.25.1 includes the following known issues.
|HIG-13622||Calendar aligned relative time syntax (last x time periods, x time periods before) is not supported in cohort definitions and time-overridden metric definitions.|
Deleting a table while a privacy purge or selective delete is in progress will cause problems.
Workaround: Wait until the purge or selective delete job has completed before attempting to delete the table.
An individual config file for a selective delete or privacy purge must be less than 16MB. However, the total allowed size for all config files that can be run at one time is 100MB.
Workaround: To run more jobs at once (than that contained in 100MB of config files), add or free up space on the data nodes. If data nodes fill up during a selective delete or purge, interrupt the jobs, verify that the query servers are all active, and then resume running half as many jobs (at once).
|HIG-13354||Cannot delete events where the column is = *null*.|
When selective delete and privacy purge interrupts are run when there are no jobs to interrupt, a success message is displayed.
Workaround: Ignore the message.
This section introduces feature enhancements included in Interana release 2.25.2, as well as newly resolved and known issues.
Release 2.25.2 includes the following enhancements.
- Support for Azure Data Lake as a data source in the Interana import pipeline.
- Privacy purge performance enhancements to improve purge completion times.
Release 2.25.2 includes the following newly resolved issues.
|HIG-14033||You were unable to set a 30-day month time resolution, the time resolution would revert to weekly.|
Release 2.25.2 includes the following known issue.
Unable to use privacy purge or selectively delete on string columns with null values.
Workaround: Contact Interana Customer Support at firstname.lastname@example.org for options to remedy this issue.