Quick one for anyone managing reload chains in Qlik Cloud: there's a Visual Task Chain Viewer now built into the Tasks section of the Analytics activity center, and it's genuinely useful.

Instead of clicking into each task in a chain to piece together what depends on what, you get a graph. Click the dependency icon next to a chained task and it opens with that task highlighted, upstream and downstream tasks laid out around it, and each connection labeled by trigger type: on succeeded, on failed, or on finished.

A few details that matter in practice:

• Task cards in the graph show the last run status, so a failed link is easy to spot.
• You can set a different task as the anchor to re-center the view, then reset back to the original.
• From a task card you can view details, edit the chain, run the task now, or activate/deactivate it, all without leaving the graph.
• Permissions still apply. If you can't access a task in the chain, you'll see the dependency but can't manage it.

I put together a full walkthrough on the Data Voyagers channel covering navigation and a couple of real troubleshooting scenarios. Link's here if you want to see it in action: https://youtu.be/GvG_wqvMlZ0

Curious if others have run into cases where this would've saved them time. Drop a comment if you've used it yet.

Regards,

Mark Costa

Hi,

We have QlikView Server and QlikView Desktop installed on a Windows Server 2022 machine, which we access via Remote Desktop.

Earlier this year, I upgraded our QlikView installation from May 2023 to September 2025 SR1. After the upgrade, I noticed that QlikView Desktop had become extremely slow and unresponsive.

When QlikView Desktop is started, the menus and the application itself initially seem to work normally. However, as soon as I open a QlikView document, the application becomes very sluggish. On AccessPoint everything works fine.

For example:

- Clicking the QlikView > About menu can take 2–4 seconds before anything happens.
- Right-clicking an object on a report can take several seconds before the context menu appears.
- Scrolling in list boxes is very jerky and noticeably delayed.
- In general, there is a significant delay between clicking something and QlikView responding.

At the time, I did not have time to investigate the issue further.

Today I uninstalled QlikView Desktop and installed the May 2023 version again. The performance immediately returned to normal. When clicking objects or menus in a QlikView document, everything responds almost instantly.

I then upgraded QlikView Desktop to May 2024 SR1, and the performance was still normal.

After that, I upgraded to September 2025 IR, and the performance problem immediately returned.

So far, my testing looks like this:

- May 2023: Good performance
- May 2024 SR1: Good performance
- September 2025 IR: Very slow / significant UI delays
- September 2025 SR1: Very slow / significant UI delays
- September 2025 SR2: Very slow / significant UI delays

I have also disabled the antivirus temporarily, but this did not make any difference.

Has anyone else experienced similar QlikView Desktop performance or UI responsiveness issues with the September 2025 releases?

I have an Oracle endpoint that I need to update with new connection information for both the Database and the ASM Log files.  Basically everything is changing  since the database is migrating from on premises Exadata to Oracle Exadata on the Cloud.  What I would like to do is create a totally new source endpoint with all the new settings ahead of time and then when the time comes to make the changes when the database moves, I would just stop the tasks, delete the old endpoint from the task and then add back the new version of that endpoint.  I seem to remember though that doing that might affect the list of tables in the task which since some of the tasks have hundreds of tables that it is dealing with.  Is that the case, or will my idea work?  The only other way I see this happening is to stop all the tasks and then update the endpoint on the fly and then restart one task that uses the endpoint to make sure it works before starting the other tasks.  

In case is matters I am running replicate 2025.11.0.770. on RHEL 8.1 on an AWS EC2 instance.  

• New Add to selection button: It replaces the search icon in the selection bar, so the first step into making a selection is easier to spot.
• Search as you type: The button opens a dropdown where you can search for dimensions, fields and values by typing.
• Selections from search results: Pick a result and add it directly to your selection, without hunting through sheets and objects first.

How to activate it
Feature previews must first be enabled for your tenant, and your tenant admin needs to grant you access via the Feature previews permission. If you don’t see Feature previews in your profile, check with your tenant admin.
Once you have access, go to your User profile > Feature previews and turn on the Add to selection preview feature. You can switch it off again at any time.
See Adding search results directly to your selection for details.

What to focus on when experimenting
We’re especially interested in hearing about:

• Whether the new button makes it clearer how to start making selections, especially for people new to Qlik
• How search feels in your real apps: result relevance, speed and ordering of dimensions, fields and values
• Whether you find what you’re looking for by typing, or still reach for other ways to select
• Anything that feels different from what you’re used to, even small things

Hi Qlik team,

We are evaluating Qlik Replicate for an on-premises CDC pipeline: Db2, SAP HANA, MySQL and PostgreSQL → Kafka → Apache Hudi. The consumer maintains current state by a stable business key and an ordering field, including logical deletes. Source versions and deployment details are still being collected.

Updated 7 October 2026 after reviewing related Community answers and the May 2026 documentation. We have narrowed the questions below to the remaining support and recovery boundaries. Answers from Qlik Support/Product specialists, with applicable release/build and configuration examples, would be very helpful.

1. Db2: supported release and licensing dependencies

The May 2026 source matrix lists LUW 11.5, while the release notes list it as end-of-life. Which supported Replicate release/SR/build should be used for a LUW 11.5 source?

We understand the different capture paths: db2ReadLog/IBM client for LUW, R4Z/IFI for z/OS, and journals for IBM i. Which additional Qlik or IBM licensing dependencies must be included for each family, and which deployment restrictions would rule out the standard CDC path?

2. SAP HANA: maintenance and failover without missing changes

We understand that trigger-based Full Record became GA in November 2025, and that May 2026 SR1 fixes Full Record before-image and resume issues.

For direct HANA log-based CDC, if Backint makes the required backup-log files inaccessible, is there a supported production alternative? For Full Record trigger mode, what supported ADD COLUMN procedure preserves the required history given the DDL restriction? The Continue CDC on missing triggers documentation says DELETE operations are not captured while triggers are absent: how should deletes during an SAP maintenance window be recovered?

Separately, can direct HANA Full Record trigger mode resume after source failover without table reload? Please distinguish this from the documented log-based failover/reload limitation.

3. Kafka: flat JSON with metadata, and tables without declared keys

The Community answer on Custom Message Format confirms section flattening. However, removing the envelope appears tied to disabling metadata publication. Can one supported configuration put business fields and selected service fields at the JSON root and also publish separate schema-change metadata messages? Please provide configuration and INSERT/UPDATE/DELETE/DDL samples, or confirm that downstream normalization is needed.

We also acknowledge that the current PostgreSQL guide supports REPLICA IDENTITY FULL, unlike older Community replies. For MySQL/PostgreSQL tables with a stable business key but no declared PK/unique index, what supported configuration preserves UPDATE/DELETE keys? We understand the CDC LOB limitation without a source PK/unique index; is there any supported alternative beyond adding a source key? Please clarify when DELETE carries a complete before-image and when the consumer must merge a delete flag with its existing row.

4. Hudi: ordering and deduplication across replay and reload

We understand at-least-once Kafka delivery and the per-partition producer ordering discussion. A January 2025 Support answer, in a Kinesis context, says that replay from an earlier timestamp resets the internal change number and changes CHANGE_SEQ for the same source record. What are the implications for our sources and Kafka consumer?

What supported event-version rule should a current-state consumer use across normal Resume, active-passive failover, replay from a supported position/timestamp, table reload and task recreation? Can an older replayed change compare greater than a newer change already applied? AR_H_CHANGE_SEQ is empty during Full Load: what non-null Hudi precombine value should initial-load records carry, and how should it compare with concurrent CDC and later reloads?

Please also identify which timestamp fields are actual database commit times for these source families, and which are Replicate-derived or may be absent. We understand that transaction IDs are not universally unique.

5. Production: durable state, disk isolation and gap-free startup

QEM RunTask now documents start-from-position, while GetTaskDetails exposes status, stop reason, latency and backlog. Is there a supported API to retrieve a durable safe resume/checkpoint position and endpoint connection health? Please specify source applicability, particularly for HANA and IBM i.

We understand that active-passive requires persistent shared DATA on supported block storage, that the critical disk threshold stops all server tasks, and that per-task resource stopping has been described as external QEM automation. Has a native bounded per-task disk queue been added? If not, what supported isolation/recovery design is recommended for Kafka outages?

The DATA-loss recovery guide lists Oracle and SQL Server targets, not Kafka. What supported recovery procedure applies to our sources with a Kafka target if DATA is lost but source logs remain? Finally, what source-specific Full Load/startup procedure prevents gaps from transactions that began before startup and outlive the consistency timeout?

Please distinguish Replicate guarantees, source/HA prerequisites and Kafka consumer responsibilities. An explicit limitation or required integration layer is useful before the pilot.

Thank you,
Alex

Qlik NPrinting May 2026 SR1 - ArgumentOutOfRange when exporting Qlik Sense Sheet as image above 800×600

Hi,

I am experiencing a reproducible issue with Qlik NPrinting May 2026 SR1 when generating an HTML report containing Qlik Sense sheets as images.

Environment

• Qlik NPrinting: May 2026 SR1

• Report type: HTML

• Source: Qlik Sense May 2026 patch 2

• Object type: Qlik Sense Sheet

• The sheets contain charts only

• No tables are used in the problematic sheets

Problem

The report fails during Content production with:

ArgumentOutOfRangeException
Заданный аргумент находится вне диапазона допустимых значений.

The NPrinting log contains:

Starting content resolution for report ...
Finished content resolution for report ...

Starting content production
WARN: error during report generation:
Заданный аргумент находится вне диапазона допустимых значений.

Finished content production

Failed generation of report ...
ERROR:
Заданный аргумент находится вне диапазона допустимых значений.

Content resolution completes successfully, and the failure happens only during content production.

Reproduction

The HTML template is extremely simple:

<html>
    <body>
        <img src="%%ef634961-f70e-4980-b27d-7a035aa28007%%" alt="Sheet 1">
        <img src="%%c19f9cd5-77cd-4037-9331-dca5115bdf7b%%" alt="Sheet 2">
    </body>
</html>

Both objects are Qlik Sense sheets added to the Images node in NPrinting Designer.

The important part is the image dimensions.

I tested the same sheet with different NPrinting image dimensions:

Image dimensions Result Below 800×600 Report generation succeeds 800×600 Report generation succeeds, but the sheet is cropped Above 800×600 Report generation fails with ArgumentOutOfRangeException Auto / 0×0 Report generation fails with ArgumentOutOfRangeException

For example, increasing the dimensions beyond 800×600 causes the report to fail even though the same sheet works with smaller dimensions.

Additional observations

The issue does not appear to be related to the HTML itself.

The same HTML report can successfully generate other content, and the failure occurs before publishing, during content production.

The problematic objects are whole Qlik Sense sheets exported as a single image.

The issue appeared after moving to the May 2026 generation of NPrinting. The same reporting approach worked in the previous NPrinting version.

According to the NPrinting documentation, Qlik Sense sheets are exported as a single image:

https://help.qlik.com/en-US/nprinting/May2026/Content/NPrinting/QlikViewEntityReports/Embed-QV-Sheet-Report-Temp.htm

Questions

1. Is there a known defect in NPrinting May 2026 / May 2026 SR1 related to Qlik Sense Sheet image rendering and image dimensions?

2. Is 800×600 an unintended rendering limit in the current version?

3. Is there a configuration parameter or workaround that allows Qlik Sense sheets to be rendered at, for example, 1920×1080?

4. Is there a recommended workaround for HTML reports when a whole Qlik Sense sheet needs to be exported as an image?

5. Could this be related to the image rendering changes/regressions introduced in May 2026?

I can provide the NPrinting logs, screenshots of the template/image properties, and a minimal Qlik Sense app reproducing the issue if required.

Thanks.