Apache Iceberg promises that you can store data once and query it with whatever engine suits the job. That's worth real money: moving Bronze and Silver transformations from Snowflake warehouses to Athena can cut their compute cost by roughly 20–45%.

The specification meets that promise, but the bill and the governance don't always follow. When Athena, Databricks, Spark or Trino read a Snowflake-managed Iceberg table, it's rarely "the engine just reads the files":

→ Row- and column-level security doesn't cross a catalog boundary for free. The two paths that enforce Snowflake's policies for an external engine both run on Snowflake compute.

→ Federation collapses identity. Glue and Unity reach Snowflake's catalog as one service principal, so per-user rules have to be declared again on the consumer side.

→ Delete files are a bigger compatibility boundary than catalogs. A table updated by an Athena MERGE can be unreadable in Databricks until it's compacted.

Capturing the savings means deciding catalog ownership tier by tier, and budgeting for the second governance estate that comes with it.

The full post has the cost model and design rules, plus a workbook and a one-query check of your Snowflake $ per TB scanned, so you can test the numbers against your own workload.

Who owns the catalog for each tier in your lakehouse, and did the numbers or the governance decide it?

Full Post URL: https://eost.dev/posts/iceberg-interoperability-has-a-bill

#ApacheIceberg #Snowflake #DataEngineering #Lakehouse #FinOps

Hello, IA is being tested by most of our customers. As everybody knows all SaaS products come with a number of free Qlik Answers questions: 25 for Starter, 200 for SaaS user based or Standard and 1.000 for Premium and Enterprise. This year Qlik changed the rule for overage. Only Capacity has overage enabled. All other features come with overage disabled, including Answers.

We came across several customer using beyound their contracted limits. So we have two questions: 1 - Who and how can we flag the tenant to disable Answers overage, 2 - Are we going to receive invoices for these overages?

Thank you 

Hi everyone,

I just published a new Data Voyagers video on Qlik Code, the Visual Studio Code extension that lets you edit Qlik Cloud load scripts in VS Code instead of the Data load editor.

I wanted to cover the whole thing in one place, so the video goes through each feature in order. Once you sign in to your tenant, you can browse spaces, apps, and scripts, then view, edit, and sync load scripts using direct or local editing. The editor supports the Qlik script language with syntax highlighting, completions, hover help, and diagnostics. You can also insert script content to load files from the data catalog or a data connection, or pull in existing script assets like QVS files. When you are ready to test, you can reload the app, preview your data, look at the data model, and open the app in Qlik Cloud from your browser.

A few limits from the documentation are worth knowing. Qlik Code is aimed at developers who already work in Qlik Cloud, and it does not replace the activity centers, so tasks like managing spaces, sheets, and app permissions stay there.

Watch the video here: https://youtu.be/8omcueKfuQ8

Documentation: https://help.qlik.com/en-US/cloud-services/Subsystems/Hub/Content/Sense_Hub/QlikCode/code-extension-home.htm

I would like to hear how this fits your workflow. Are you already writing scripts in VS Code, or still in the Data load editor? Which feature would you use first, and what is missing for you? Tell me in the comments.

Regards,

Mark Costa

I wanted to know if there was an implementation plan for the Semantic Model.  Creating a new layer feels like a rather large project and it would be nice if there was some guidance available on how to go about it systematically.

How much of our existing Qlik application metadata and master measures can the semantic model automatically reuse, and how much must we manually recreate?

If it's mostly manual, I'd want a demonstrated business benefit before agreeing to implement it.

I have a table that currently works - it allows the user to select WeekStartShipDate, which is loaded in the load script with a Floor value, filtering the table to just one week:

The variables vStartSixWk and vEndSixWk determine the date range for the 6 Week column. 

vStartSixWk  =(WeekStartShipDate) -42

vEndSixWk  =(WeekStartShipDate)

The 6 Week column is then calculated as follows:

Sum({<WeekEndShipDate={">=$(vStartSixWk)<=$(vEndSixWk)"}>} EquivalentUnits) / 6

The % of 6Wk Total column is calculated as follows:

Num(
Sum({<WeekEndShipDate={">=$(vStartSixWk)<=$(vEndSixWk)"}>} EquivalentUnits) / 6
/
Sum(
TOTAL
Aggr(
Sum({<WeekEndShipDate={">=$(vStartSixWk)<=$(vEndSixWk)"}>} EquivalentUnits) / 6,
ShipTo
)
),
'##0.0%'
)

So, that all works fine until I add another dimension column, such as SalesOrder. Once I do that, the date filter breaks, and the 6 Week expressions return zeroes.

Please advise.

Hi everyone,

We are experiencing an OAuth 2.0 authentication issue when trying to connect Qlik Cloud Analytics to Salesforce using the Salesforce connector from Add Data.

A new Salesforce External Client App was created following Salesforce's recent security changes.

According to Salesforce's updated security requirements, PKCE is mandatory for newly created applications, and the option to disable it is locked.

However, the Qlik documentation instructs us to disable PKCE as part of the connection setup.

We attempted to configure the Salesforce Callback URL using both:

https://connector.qlik.com/auth/oauth/v2.htm

https://connector.qlik.com/auth/oauth/v3.htm

Both attempts failed during authentication, resulting in the following errors:

redirect_uri_mismatch

OAUTH_APPROVAL_ERROR_GENERIC

We would appreciate hearing from anyone who has encountered a similar issue or has successfully established a Salesforce OAuth 2.0 connection in Qlik Cloud Analytics under the new PKCE requirements.

Thanks in advance for sharing your experience!

I'm  exploring this topic and finding some interesting way of enhancing our Qlik viz experience

SVG in Qlik 

One qvf attached one to download from the website.

Enjoy and Qlik:)

This version adds support for migrating of Qlik Sense Task chains, adds an option to use NTLM Authentication for the Service account and introduces Silent install.

Task chains are displayed in a new Task tab in the Plan, allowing to edit the individual task or chain with options to split an invalid or unnecessary chain, attaching specific tasks to new or existing ones and other types of fixes.

The Plan page tables have been enhanced to show warnings and other information, and have been grouped depending on the source, Qlik Sense, QlikView or Qlik NPrinting.

Using NTLM for the authentication of the Service account removes the need to set a Service Principal Name (SPN), which prevented a successful installation on servers that already had somme other service already running. Other methods of authentication are still supported.

It is now possible to perform a silent install based on a file containing all the information required for the installation, allowing for prgramatic installation.

Qlik Analytics Migration Tool overview

Click here to download

Our infrastructure security scanner flagged our Qlik Sense servers as potentially vulnerable because of the bundled Node.js 24.18.1 runtime. Following the October 21, 2026 Node.js security release, I tested updating that binary independently of a Qlik patch.

A scanner finding against a runtime version does not establish that the vulnerabilities are exploitable through Qlik Sense. Product-specific impact still needs confirmation from Qlik.

The update from Node.js 24.18.1 to 24.22.1 passed testing in our Qlik Sense Enterprise on Windows May 2026 Patch 7 environment. This is one environment's test result, not a compatibility guarantee.

This experimental method is not validated by Qlik. Qlik's documented approach is to update the bundled runtime through a Qlik Sense upgrade. This post shares a technical test, not an official remediation recommendation. If evaluating it, use development first, a maintenance window and a recoverable full backup.

Check the bundled version

& "$env:ProgramFiles\Qlik\Sense\ServiceDispatcher\Node\node.exe" --version

The full path checks Qlik's runtime rather than another Node.js installation on the PATH.

Run the attached scripts

1. Extract the scripts into one folder.
2. Download the official Node.js 24.22.1 Windows x64 binary.
3. Edit NodeUpdate.config.psd1: configure the binary, backup and log paths, and set the expected SHA256 using the win-x64/node.exe entry in SHASUMS256.txt.
4. Open Windows PowerShell 5.1 as administrator and run the backup:

.\01_Backup_Node.ps1

Continue only after BACKUP VERIFIED:

.\02_Update_Node.ps1

Leave the Service Dispatcher running before starting the update: the script handles stopping and restarting it. Execution is local to each machine; relative paths are resolved against the configuration folder.

Validate and roll back

Check Hub, QMC, application access, exports, reloads, usual connectors and Qlik logs. These are suggested checks, not a claim that all were covered in our test.

To roll back, set RestoreBackupDirectory to the exact original backup folder, then run:

.\03_Restore_Node.ps1

The scripts target the main Service Dispatcher runtime on Node.js 24 only. They do not update EAPowerTools, Package Cache or npm dependencies. A binary backup is not a full site backup, and automatic rollback is not guaranteed. Recheck the installed version after subsequent Qlik maintenance.

Sources: Node.js security bulletin, October 21, 2026; Node.js 24.22.1 release notes, October 21, 2026; Qlik Support guidance, updated August 14, 2024.

Hi @Øystein_Kolsrud ,

I'm trying to use Qlik Explorer for Developers to analyze application opening performance (Cold and Warm Open) in Qlik Sense Enterprise on Windows.

However, the Progress Report always shows "No progress data available", even though the application opens successfully.

Environment:

• Qlik Explorer for Developers: 2.11.1

• Qlik Sense Enterprise on Windows November 2025 p15 and May 2026 P7

• Tested both through an Nginx reverse proxy and with a direct connection to Qlik Sense

• info template selected

I enabled WebSocket traffic logging and found something interesting.

The OpenDoc operation completes successfully, and the Engine returns detailed progress information through GetProgress.

For example, during a Cold Open test:

• Loading from storage

• Loading fields

• Loading tables

• Attached to loaded document

• Open document end, time 1794.8ms

In this test, 34 progress messages were successfully received across multiple GetProgress calls, including the final response with qFinished: true.

However, none of this information appears in the Progress Report section of Qlik Explorer.

The same behavior occurs during Warm Open, and I reproduced it both with and without the external reverse proxy.

This suggests that the Engine provides the required information, but Qlik Explorer may not be displaying or processing it correctly.

Is this a known issue or am I missing some configuration?

I can share sanitized WebSocket traffic logs if needed.

Thanks!