HQ
Change log for Lenses HQ 6.2.
6.2.4
Improvements
IAM resources, Audit and Alert creation is now formed based for consistency.
Kafka Connections are now organized by the environment they connect to for better clarity.
SQL Processor creation flow has a separate tab for SQL to provide more space on smaller screens.
Favourites resources now inherit the environment colour.
Alert Event layout has been improved, now the details opens in a sub panel to the right, matching the data explorer row experience for consistency.
Search in schema on topic listing tab now shows a count of matches per topic.
SQL Processors now support accessing the metadata for records on simple selects and lateral joins. e.g.
SELECT _meta.partition/offset/timestamp/topic.No support yet for join / group by. Requires Agent upgrade.Accept terms and conditions in community edition can now be accepted from none admin users.
Fixed
Empty leaf breadcrumb node for ACL creation has been fixed.
Empty in-sync replicas chip for topics partitions has been fixed.
Fixed a bug that stopped the details showing for search in schema on topic listing page.
Inserting auto-generated samples into AVRO topics with
decimal/bytes/fixedfields.
6.2.3
Download data
You can now download SQL query results as CSV or JSON.
Data Policies now with dedicated Impacts tab
Data policies now have a dedicated tab for impacts rather than the information panels only.
Tab Context Menu Actions more visible
Tab context menu actions are now clearer and moved into the tab content from the breadcrumb row. These actions are context aware allowing the user to perform relevant actions or view different parts of the a resource, for example, insert messages into a topic or view its configuration.
Improvements
Form creation. Creation of Data Policies, Kafka ACLs and Quotas are not form based for consistency
Kafka Connectors. Kafka Connector creation problem panel has been updated from two separate tabs into one for clear identification of validation issues from Connects API and Schema issues.
Agent configuration. The provisioning file icon for agent configuration testing has moved up into a dedicated button making it more visible.
K2K. Kafka to Kafka now has a force delete option to ensure dangling Kubernetes resources are clean up correctly.
Improved error guidance in K2K Replicator setup
Error messages displayed during Kafka-to-Kafka (K2K) replicator creation now include actionable links to help you resolve configuration issues faster:
"No matching deployment found" — This error has been upgraded to a more visible alert and now includes a direct link to the Kafka Connections documentation as well as a "Create or manage Kafka Connections" action that opens the Connections listing in a side-by-side tab, so you can fix the issue without losing context in the K2K form.
"No Kubernetes deployment targets found" — This warning now links directly to the Agent Helm deployment documentation, guiding you through the required Kubernetes cluster configuration.
These changes eliminate the guesswork previously required when encountering deployment errors during K2K setup.
Fixed
Breadcrumbs. Create topic breadcrumb now shows the correct levels.
SQL Queries. SQL Queries losing syntax highlighting has now been fixed.
Auto-completion. Fixed OS commands for triggering auto-completion in editors, they now correctly show Ctrl+space on MacOS instead of Cmd+Space.
App navigation tree. Fixed a issue for Kafka Connector, SQL Processor and External Apps navigation trees showing an expand chevron for Connect cluster with no connectors.
Kafka ACLs. Fixed an issue that prevented the creation and listing of Kafka ACLs with more than one
:separator in the principal for SPIFFE format.Navigation Tree. Fixed an issue with the navigation tree now allowing scrolling on very small window sizes.
Schema Creation. Fixed an issue that caused schema creation to fail when primitive types where used for a key schema.
Navigation warning. Don't warn the user about unsaved tabs when leaving studio to go to organization settings if no tab is dirty.
Filter drop downs. Fixed a bug where filter state was not maintained when changing tab focus.
Listing open multiple tabs. Fixed a bug that duplicated opening listing tabs. Now if the tab is already open focus will be set to it and only opened a new tab if not already open.
K2K Save & Restart Button. Fixed a bug that prevented the enabling the save and restart button during editing a K2K application.
Data Policies. Fixed a bug that stopped user providing their own categories during creation.
6.2.2
Single config editor for Kafka Connector creation
Creating a Kafka Connector now uses one unified configuration editor instead of separate tabs that could silently lose keys when you switched between them. Code completions work across the whole config in one place.
K2K Health: stale data indicator during agent degradation
The K2K Health page now continues to show source and destination partition data when a Lenses agent is temporarily unavailable — for example, when the agent is timing out on a large Kafka cluster or is restarting. Previously, any error communicating with an agent would cause the source and destination partition columns to go blank for the duration of the outage, with no indication that the displayed data was stale.
Source and destination partition data is preserved from cache when an agent fails to respond. Lag, consumer offsets, and log-end offsets remain visible.
A "Stale data" chip appears on the Health card whenever partition data is served from cache, and independently on the Throughput card when consumer offset metrics are stale.
Partition data is shown as N/A (rather than blank) when the cache is cold and no prior data is available.
Topic name clicks open the topic profile directly instead of a SQL query tab.
Improvements
Persistent tab state across more create/edit flows. Work-in-progress in the remaining Studio create/edit tabs now survives tab switches.
BYTES topic warning links to change format type. The SQL engine uses an internal type to de-serialize records. When a topic is set to BYTES, the warning in SQL Studio now includes a direct link to the change-format-type modal, so you don't have to dig for it.
Create connector from topic context menu. From the topic context menu in the nav tree, a shortcut now creates a Kafka Connector for that environment.
Connector metrics cards adapt to narrow viewports. Metrics cards on the connector page wrap correctly when the window is narrow.
Nav tree reorganization. The navigation tree has been reorganized for clarity. Environment adminstation is now under an admin node.
Agent key is front-and-center during environment creation. The agent key is now clearly visible for you to copy — along with the command to regenerate it — instead of being buried in the Docker and Helm install snippets.
Fixed
Topic Partitions tab no longer crashes. Fixed an infinite re-render loop that was crashing the Topic Partitions tab.
SQL query tab layout. Fixed a bad layout on the SQL query tab.
Sign-in loop. Returning users in certain auth states could get stuck in a sign-in loop; this is fixed.
Connector breadcrumbs replace the current tab. Clicking a connector breadcrumb now replaces the current tab rather than opening a new one.
Audit filter race condition. A race condition when fetching audit entries left some filters under-populated. Filters now correctly include all values.
"Copy and close" closes the MCP setup modal on landing page. The Copy-and-close action in the MCP setup modal now actually closes the modal.
Explorer bar icons wrap on small screens. Explorer bar icons (Home, Environment, All topics, Favorites) now wrap on resize so they stay visible instead of being clipped.
6.2.1
New Studio welcome experience
The post-login welcome modal is gone. The Studio welcome tab is now your real landing surface — a single onboarding hub that orients new users and gives experienced ones one-click entry into every part of Lenses.
At the top of the tab, short Environments and Agents primers put the mental model in front of you before you start clicking, alongside an embedded quickstart video that you can dismiss or open in a new tab.
Below that, a search across environments, topics, applications and AI entry points sits above a curated action grid grouped into six themes:
Connect to your cluster — create an Environment and install the VS Code, Cursor, IBM Bob, or Antigravity extensions.
AI & IDE extensions — connect via MCP to onboard your AI assistants.
Discover & query — view environments, explore topics, query topics with SQL, create topics, create schemas.
Process & replicate — SQL Processors, Kafka Connectors, K2K, App Migrations, External Apps.
Governance — IAM Users, Groups, Service Accounts, Roles.
Resources — quickstart video, documentation, the Lenses Slack community, Ask Marios community Q&A, and a direct line to support.
Two guided wizards live behind the relevant cards: a MCP connection setup that walks you through Docker plus client-specific config and CLI, with OAuth-vs-API-key detection from your settings and a service-account shortcut, and a guided IDE extension install with deep-links and copyable CLI snippets for VS Code, Cursor, IBM Bob, and Antigravity.
Open in VS Code from more places
Following the VS Code Plugin introduced in 6.2.0, the "Open in VS Code" deep link is now wired up across all supported screens — including a direct link from SQL Studio. Pair it with the Lenses VS Code Plugin to jump from HQ straight into your IDE without breaking your flow.
Improvements
Studio tab state persists across sessions. When you reopen Studio, your tabs come back the way you left them.
Inline editing for topic profile types. Key and value types on the topic profile page can be edited inline, with edit icons next to each field. The same flow surfaces tags inline as well.
Consistent JSON grid view. Studio query results and the topic data tab now use the same JSON grid view and message details, so your eyes don't have to switch contexts.
Large nested SQL results no longer freeze the browser. Studio handles deeply nested results gracefully.
Empty-state Create button in the SQL Processors Studio tab now actually creates a processor.
Long topic names truncate with an ellipsis and full-name tooltip in the explorer.
Search panel state is preserved when you switch between sidebar panels — no more retyping your last query.
K2K Save vs Deploy is no longer ambiguous. The K2K create flow uses a split button so it's clear whether you're saving the configuration or deploying the application.
K2K "Start", not "Resume", for new K2K apps. Apps that have never run no longer show a misleading "Resume" button.
K2K Target environment name in info panels. The K2K information panel now shows the name of the target environment.
App Migration shows source and target offset mapping. Before resuming, you can see where each topic's consumer group is and where it's about to land on the target cluster.
App migration context menu correctly populates the consumer group when launched from the consumer-group screen.
Auto-converter selection, restart context menu, and dark-mode icons in connectors all behave correctly.
Configure a disconnected agent. You can now reach an agent's configuration page even when the agent is disconnected, so you can prepare provisioning before bringing it online.
Warning when JMX is not configured. If JMX is missing for one or more services, Environment Health surfaces a warning rather than silently degrading the dashboard.
Auto-expand the nav tree when only one environment exists.
Lenses Resource Names (LRNs) are now shown in information panels.
Data policy moved into the Studio information panel. The standalone side drawer is gone; everything lives next to the resource it describes.
Connector create config editor is more stable under heavy autocompletion: cursor position is preserved, multi-word autocomplete works, and topic names are validated as you type.
Connector restart context menu, dark-mode icons, and styling for connectors are fixed; the nav tree refreshes after a connector is created.
OAuth 2.1: Clearer errors on Dynamic Client Registration. When a client registration fails, the response now explains why instead of returning a generic error.
OAuth 2.1: Bulk delete for registered clients. Clean up many clients in a single call instead of one-by-one.
Fewer re-renders on busy screens thanks to wider use of
React.memoon high-traffic components — noticeable on screens with many resources.MCP setup snippets for Helm use a
[LENSES_HQ_URL]placeholder so the snippet is unambiguous when you copy-paste.Admin panel padding and overflow improvements.
Fixed
Debug-mode telemetry removed. Telemetry that ran in debug mode could expose environment details; it has been removed.
SSO redirect URL validation tightened to reject malformed or hostile URLs.
Deleting an already-deleted topic no longer freezes the UI.
No more "white box" when a topic is deleted from the listing.
Partition page updates correctly when a topic is recreated empty.
Bulk topic delete uses the bulk endpoint and the modal closes correctly when topics fail to delete.
Brief error screen on initial app load before the login redirect is fixed.
Loading flicker in Studio when no Schema Registry is connected is fixed.
Schema Registry node in the nav tree correctly reflects whether you're connected.
K2K per-partition metrics display correctly.
K2K throughput metrics are reported at the topic level.
K2K license issue is no longer wrongfully shown.
K2K broker count is correct when a cluster is in a degraded state.
Empty modal after consumer-group deletion no longer blocks the UI.
Dark-mode icons for connectors and other surfaces are correct.
UI clipping on standard laptop resolutions, including the missing scroll on the create-environment deploy step, is fixed.
Counts on listing pages wrap correctly on smaller windows.
6.2.0
SQL Studio Becomes Studio
We’ve totally revamped the global SQL Studio and renamed it to just Studio. It’s designed to deliver maximum productivity and simplicity for executing engineering workflows. The experience (including shortcuts) is inspired by your favourite IDEs.
The previous HQ views (Apps, IAM, …) in 6.0 and 6.1 have been integrated into this Studio as well as rolling up more assets from the legacy UI.
This new Studio provides a unified experience to manage all Kafka and Lenses resources including Lenses IAM roles, SQL Processors, Connectors, K2K apps, Kafka Connections, Schemas, Topics, Consumers and more.
The Studio also comes with a new search to easily find any resources across your multi-Kafka estate.
The Environments view, accessible from the top of the Explorer, has been updated to centrally view the health snapshot of your different connected Kafka environments.
Access to the global Topics catalog has also been moved to the top of the Explorer.
The Bookmarks in 6.1 have been renamed to Favourites where any Kafka or Lenses resource can be saved for easy access, avoiding the need to search or navigate the Tree Explorer.
A Form/YAML toggle will be noticed on a number of different pages – allowing users to select their preferred configuration style.
SQL Processors in HQ
SQL Processors now have a new experience in HQ, giving a global location to manage and create stream processing applications.
Listing:
View and search all SQL Processors across all environments
Bulk Action support to stop, start and delete processors
To create SQL Processors:
Improved editor, with better intelli-sense, including formats, partitions, replicas and topic size and message stats
Templates for common SQL Processing patterns
Auto task max, that will automatically set the number of SQL Processor runners based on the topics involved
SQL Processor details:
More information is surfaced to allow you to quickly see:
Health information: Throughput & message counts, errors and runner logs
Consumer groups, lag and the ability to manage those offsets
Alerts: Any alerts raised for this consumer or output topics and ability to create new alerts directly from the SQL Processor page
Interactive topology of the processor to understand visually what it’s doing.
Connectors in HQ
Kafka Connectors also now have a new experience in HQ. The capabilities are extensive, providing:
JSON schema support to ensure you minimise configuration mistakes
Auto completion examples for KCQL for Stream Reactor connectors
Auto task max, making sure you correctly scale your connectors based on topic partitions
Automatic converter selection based on topic format
Auto completion default snippets for
Converters
Producer and Consumer overrides
Error handling
Support resetting Connector offsets, source and sink.
Requires Kafka Connect 3.7 and above
External Apps in HQ
Last but not least, registering external applications into the Lenses App Catalog and Topology is also in HQ.
In the coming releases, we’ll be updating Topology to HQ too, to give you a global view of application and data lineage across Kafka clusters.
Secure Remote MCP Server
You loved the Lenses MCP Server released last year. But now you are looking to roll it out to more Agents and users. For this, simplicity, scalability and security are critical to you.
Now in this release, rather than having to deploy the Lenses MCP Server locally across machines and workstations, we’ve included it in the Lenses deployment running as a server. This allows local MCP Clients such as Cursor or Claude to securely connect to Lenses via MCP, without having to roll out multiple Servers.
Instead of using STDIO, the Server uses a single Streaming HTTP endpoint, as recommended by the MCP community.
And perhaps most importantly, it also allows the support of OAuth 2.1 authorization to the MCP Server (see below).
OAuth 2.1 Support
Handing an agent a long-lived API key isn't defensible. The Lenses MCP Server now supports OAuth 2.1 for delegated access, replacing static credentials with short-lived tokens tied to a named user.
Three scopes ( lenses:read, lenses:write, lenses:delete ) sit on top of the user's existing Lenses IAM role and let the agent inherit a subset of the user's IAM permissions, capped by what they consented to.
Access tokens are short-lived (minutes, not days) and refresh transparently in the background, so users aren't prompted to sign in mid-session. Every call is attributable to the user who authorized it in the audit trail, and any registered client can be revoked from within Lenses in seconds.
VSCode Plugin
You love the Lenses UI but you also often want to execute some workflows without leaving your IDE.
Now you can – we are pleased to offer the experience of Lenses within a VSCode-compatible IDE including Microsoft Visual Studio Code and Cursor.
This first release has focused on allowing users to configure Environments (connect Lenses to remote Kafka clusters), explore topics with Lenses Snapshot SQL, view and manage schemas and consumer groups, and manage Lenses IAM users and roles as-code.
But the VS Code Plugin is on a different release track than Lenses, so keep a close eye for very regular updates. Check-out the separate docs.
The VS Code Plugin can be downloaded and installed from the Microsoft VS Code Marketplace.
Kafka App Migrations
Replicating data from one Kafka cluster to another with K2K is great, but what about the consumer groups?
The committed offsets don’t match between source and target clusters, so if you blindly re-point your consumer to the new cluster, it might reprocess data.
This might be ok for your use case, but likely not.
We know that migrations can take months if not years if you have hundreds or thousands of apps to migrate, and it comes with a lot of risk and complexity.
This has now been drastically simplified thanks to the visibility and control that Lenses now offers, including a new “Migration” workflow within the Apps section of Lenses that allows you to migrate and translate the consumer group offsets to the target cluster with a single click!
Schema Inference
For topics not serialised through a schema in a Schema Registry (such as CSV or JSON topics), Lenses infers the schema. But these schemas could change over time, which breaks how Lenses is able to view the data.
In these circumstances you can now trigger schema inference on the Schema tab within Studio for a topic.
Improvements
K2K Apps
Some of you like YAML, some don’t — we understand. The Kafka 2 Kafka creation and edit workflows have been updated to include a more traditional approach. You can either use the form to create a replicator, or go hardcore and use the YAML editor with auto completion and intelli-sense. Pick your poison.
We also surfaced more information in the listing side drawer and details pages, improving the layout and we also added bulk actions to pause, start and delete applications.
You also now get an interactive topology of your K2K application, enabling you to perform actions and drill into the topics involved.
Consumer Group Offset Management
K2K applications now expose full control over their consumer group offsets to replay or control data to be replicated to the target cluster. You can view where replication stands – per topic, per partition – and reset it without touching Kafka directly.
From the new Consumer Groups screen you can:
View committed offsets for all topics in the application’s consumer group
Reset offsets by timestamp — Lenses resolves the closest available offset per partition automatically
Reset offsets to a specific raw value (e.g. 0 to replay from the beginning)
Operate at any scope: the entire application, a selection of topics, or a single partition
The application must be paused before offsets can be modified; all controls are disabled while it is running.
When creating a new K2K application you can also now set startFrom to control where replication begins: earliest to replay all history, latest for new messages only, or an RFC3339 timestamp to start from a specific point in time.
Surface Error states
We have improved the error reporting to give more visibility into the state of the application.
Lenses Agent Configuration
Deploying and configuring a Lenses Agent to a Kafka environment now includes more guidance on uploading JKS files when mTLS is required, and is now managed in the new Studio in HQ. The editor will also mark the gutter when a file needs to be uploaded.
Hints to change data format in SQL Snapshot
The Lenses SQL Snapshot engine tries to determine the data format in topics. It doesn’t change the underlying data; this is just metadata in Lenses. If it can’t determine the format, it defaults to BYTES. You can control this via the content menus to change the type.
If you query data with SQL with the type set to BYTES, the records are shown in the bad records tab. We have enhanced this to let you change the type, with guidance to help you choose.
Inserting messages
In Studio, you can now choose which partition to insert the message into.
Bulk Actions for Topics and Schemas
You can now perform bulk actions such as delete on multiple topics or schemas.
Info panels
On the right-hand side drawer, a per-resource-type information panel is available.
Environments
You can now go to the configuration section if your agent is disconnected. This allows you to configure the provisioning and apply to the agent at deployment time or via CI/CD.
Kafka ACLs
Filtering by operation has been added.
Fixed
Applying tags to topics is now fixed.
Increasing partitions for a topic no longer results in a data format error.
Deletion of SQL Processors in a failed state now works correctly.
Last updated
Was this helpful?