Service Pack 4

Release date: May 14, 2026

Service pack availability

Service Pack 4 is available for the following release families:

  • 2025.7: Release 2025.7.8
  • 2026.3: Release 2026.3.2
    Important: 2026.3.2 Data Agent supported version

    A new data agent version (14.0.0) is required for the 2026.3.2 release to incorporate the fixes introduced in this service pack.

    For details about how to upgrade the data agent, refer to the Data Agent Upgrade considerations in the 2026.3.0 release notes.

Recommendation

It is strongly recommended to upgrade to the latest available service pack within your release family to ensure optimal stability, security, and continued support coverage.

Fixed issues

Bug IdentifierSupport TicketDescriptionApplicable Releases
INC-10031665443Dashboards did not render due to a locking issue that blocked processing.2025.7.8
INC-99469
INC-100784
Incorta Intelligence queries timed out before Advanced SQLi could complete the query, while Advanced SQLi continued running in the background, consuming resources. The fix ensures that Advanced SQLi now respects the Incorta Intelligence timeout configuration.2025.7.8
INC-10029365378Intermittent disconnection issues when querying data in Power BI via Advanced SQLi.2025.7.8
INC-10079766201Incremental loads on parent tables caused joins to be calculated incrementally, leading to potential inaccuracies when using FIRST_VERSION or LAST_VERSION in join filters. This fix ensures full join recalculation in these cases.2025.7.8
INC-10056165462Changing a materialized view (MV) to an SQL data source caused the data source to appear blank and prevented the table from loading after saving the changes.2025.7.8
INC-100375
INC-100376
66040
65586
The Spark property spark.dynamicAllocation.maxExecutors in the CMC or spark-defaults.conf was previously ignored, while it was only effective when configured at the individual materialized view level.2025.7.8
INC-10038765762The intensive cleanup job incorrectly handled tables that were created after the cleanup process had already started.2025.7.8
INC-10079265940Applying measure filters on aggregate functions using groupBy(), such as max(..., groupBy()), caused errors in tabular insights.2025.7.8
INC-10084264917Cleanup jobs incorrectly deleted compacted versions for non-optimized tables as they lack a committed version. The fix ensures using the recent version instead of the committed version as the deletion threshold.2025.7.8
2026.3.2
INC-10041865547The dashboard history cleanup job scheduling process blocked the Analytics service startup. The fix moves this process to a background thread, allowing the Analytics service to start up without waiting for it to complete.2025.7.8
2026.3.2
INC-10095066547A hang occurred when using secure headers with Automatic Web UI Updates. Analytics service reliability is now improved.2026.3.2
INC-10106466820In some cases, tagging a user with @mention in dashboard chat comments incorrectly displayed the user as @deleted instead of the actual username.2026.3.2
INC-10063965840Text in transposed Aggregated tables was not fully displayed.2026.3.2
INC-101052Charts did not support controlling the spacing between data point groups. A configurable grouping gutter has been added.2026.3.2
INC-9913963336When formatting measure headers in a pivot table with Dynamic Group-by enabled, the first column was incorrectly treated as a corner cell rather than a column header.2026.3.2
INC-10031665443Dashboards failed to load with The underlying in this insight is being updated error, even when no loads or edits were running.2026.3.2
INC-10079566297In Actual Size mode, the Growth KPI line was missing for Windows users while displaying correctly on Mac.2026.3.2
INC-10079265940Applying a measure filter on a formula using aggregate functions with groupBy() caused an error.2026.3.2
INC-10076866240The dashboard filter menu appeared differently on Safari than on Chrome when using an iPad.2026.3.2
INC-10123467027The generation of the data agent authorization file failed after upgrading from release 2024.7.6 or later in the 2024.7.x family to later releases.2026.3.2
INC-10096366319Insights built on PostgreSQL views returned a INC_06020101: SQLi Error – Populating shared temp table failed error.2026.3.2
INC-101164When the planning phase failed during the transformation node generation, the error message did not indicate which table failed.2026.3.2
INC-10101266643Running concurrent API requests using the same personal access token (PAT) caused response latency.2026.3.2
INC-10095066547The Analytics service was not responding because the Tomcat Thread Pool was consumed, requiring restarting all Analytics services. The fix ensures supporting the interruption of blocking processes, allowing system recovery.2026.3.2
INC-10038265694When the Data Agent failed to auto-upgrade due to connectivity issues, such as blocked ports, the error message was unclear. The fix improves error reporting.2026.3.2
INC-10036165681When using the Data Agent, incremental loads that extracted zero rows incorrectly reset the maximum value incremental reference to zero instead of preserving the previous value, causing subsequent incremental loads to extract the full data.2026.3.2
INC-10085466494In multi-tenant environments, the Data Agent Controller simultaneously initialized multiple agent processes per authentication file, causing a race condition where multiple services attempted to start the same data agent concurrently. The fix limits the DataAgentRestartService to be only one per data agent not per tenant.2026.3.2
INC-10091566948
66596
Intermittent long-running and failed Advanced SQLi (Kyuubi) connections caused dashboard insights to fail with a JDBC connection reset error. The fix improves connection timeout handling and stale connection cleanup.2026.3.2
INC-9954964367The Analytics audit CSV file was created as an empty file, causing subsequent table load jobs to fail. The root cause could not be identified. The fix adds error logging in the Auditor in case an exception occurs when writing audit files.2026.3.2
INC-95707
INC-96164
60498Materialized Views did not trim string values before writing to Parquet, unlike Incorta's native Parquet writer, causing string comparison inconsistencies between regular and MV-produced tables. The fix introduces a new configurable property (spark.preprocess.string.trim) to align MV string handling with the native Parquet writer behavior.2026.3.2