Feature Releases and Improvements

Q3 2025

• Topic Browser Live Updates
  • Topics now highlight in real time when they receive new data — just like MQTT Explorer.
  • A brief animation appears on the entire path whenever values update.
  • Makes it easy to spot active data flows without manually checking each topic.

• Dynamic Location Levels
  • Define custom hierarchical location levels for bridges and stream processors.
  • Create as many levels as needed (Level 0, Level 1, Level 2, etc.) to fit your structure.
  • The first level (Level 0) is always required and visible for consistency.
• Enhanced User Permissions (V2 Certificates)
  • New certificate system with granular, location-based roles.
  • Assign specific permissions per level — from enterprise to line.
  • Fully redesigned UI with better forms, validation, and clarity.
• Central access management for UNS
  • Unified access control for your entire Unified Namespace.
  • Single login for all authorized systems, with ISA-95-level permissions.
  • Integrates directly with Active Directory, SAML 2.0, and MFA.
  • No VPN chains or jump hosts — production data stays securely on your network.

• New benthos-umh version
  • Adds sanitized SparkplugB metadata fields automatically.
  • Simplifies UMH topic creation in user-defined processing.
  • Eliminates manual cleanup of special characters from IDs.
• Automatic Resource Protection
  • (Disabled by default) Prevents too many data connections when CPU, memory, or disk are limited.
• Automatically pauses and resumes connections as resources recover.
• Disabled by default to allow full control.
• CPU Throttling Detection
  • Detects when CPU throttling occurs due to resource constraints.
  • Improves visibility into performance limits during heavy workloads.

Improvements

• Redesigned Bridge Configuration - We've completely reorganized the bridge configuration interface for better clarity:,
  • Split into three logical tabs: General, Connection, and Read Flow,
  • Connection details are always visible in the Read Flow tab, so you always know which bridge you're configuring,
  • New bridges now properly enable all required tabs with clear "Required" badges,
  • Bridge monitoring shows clearer status with "Active" or "Down" states plus real-time latency,
  • See also: https://supercut.ai/share/umh/EhHkzOSEO8ES4NIt9qXUJx,
• Simplified Instance Creation - UMH Core is now the default installation type for new instances, streamlining the onboarding process. The outdated requirements panel has been removed, with system requirements now linking directly to the documentation.,
• Clearer Validation Messages - Removed confusing RFC 1123 references from validation errors. Messages now simply state "Only lowercase letters, numbers, and dashes are allowed" for better clarity.,
• Automatic Report-by-Exception for All Bridges - All timeseries bridges now include an automatic downsampler that uses report-by-exception: data is only sent to the UNS when the value or payload changes. This dramatically reduces storage and network usage while preserving all meaningful data. The default configuration removes exact duplicates and provides a 30-minute heartbeat for sensor health monitoring.⚠️ Important for Testing: The downsampler filters messages with identical payloads. If you're testing with a Generate protocol or static test data (like root = "hello world" repeatedly), you'll see metrics showing messages received but not appearing in the Topic Browser. This is intended behavior - the downsampler is filtering duplicates to save storage.For testing or debugging:
  • Add msg.meta.ds_ignore = "testing" to your mapping to bypass the downsampler
  • Or use changing values in your test data: root = "test-" + uuid_v4()
  • For production use with noisy sensors, adjust the threshold to filter sensor noise (see downsampler documentation)
• Clearer Status Messages - All status messages now use plain language instead of technical jargon. What was "DFC" is now "flow", "Protocol Converter" is now "bridge", and "Source/Sink Flow" is now "Read/Write Flow". You'll see messages like "starting: waiting for flow" instead of "starting: waiting for DFC". All your existing configurations continue working—only the messages you see have changed.
• Network Reliability - Resolved "instance offline" issues caused by network instability and connection resets to management.umh.app. Added intelligent HTTP retry logic that automatically recovers from transient failures (connection resets, timeouts, EOF errors) using exponential backoff with up to 10 retries. The communicator now handles message queue overflow scenarios that previously required manual restarts. Implementation includes refactored HTTP client package for better maintainability.
• Data Processing - Updated benthos-umh to v0.11.3, adding pre-sanitized SparkplugB metadata fields (spb_group_id_sanitized, spb_edge_node_id_sanitized, etc.) that simplify UMH topic construction by providing alphanumeric versions with special characters already replaced.
• Smarter Instance Selection - The system no longer auto-selects the first instance when creating bridges, standalone flows, or stream processors. This prevents accidental deployments to the wrong instance. You'll now explicitly choose your target instance, with helpful error messages if you forget to select one.,
• Better Navigation - Click on throughput values for any Data Flow Component (bridges, standalone flows, protocol converters) to quickly navigate to their configuration pages. Previously this only worked for bridges.,
• More Robust Deployments - Fixed the "black dialog" issue that sometimes appeared when deploying standalone flows. The diff view now preloads in the background for increased reliability, ensuring the deployment dialog always displays correctly.,
• Localhost Warning - Added a warning when the user selects localhost to notify the user that he will be working iwhtin the contaienr and not on the host system

Q2 2025

• UMH-Core v0.37.0
  • Introduces the Topic Browser service (browse Kafka topics from the console), richer Redpanda health/metrics, mock-manager for system-state testing, and automatic troubleshooting docs. Also bumps benthos-umh to 0.9.4 and adds log-rotation-aware GetLogs.
• UMH-Core v0.36.0
  • Adds the first Protocol Converter service & FSM, a generic get-metrics API, faster health checks, benthos-umh 0.9.2 (with UNS plugins), and lots of reliability fixes.
  • Upgrade via the Management Console to unlock the new converter flow.
    
• UMH Core - Coming soon
  • Zero Kubernetes requirement — perfect for locked-down enterprise laptops or edge PCs.
  • Built-in buffering with Redpanda, benthos (with over 50+ industrial connectors!), and the compliance features of the Management Console (Audit Trails, SSO, etc.)
  • Runs happily on 2 vCPU / 4 GB RAM

• Native Ethernet/IP connectivity in benthos-umh
  • Harvest live tag data from CompactLogix, ControlLogix, Micro800 (and more) and stream it directly through UMH’s high-throughput pipeline.
  • One YAML file, one Docker run, and your plant telemetry is real-time-ready.

Learn more about this feature in this dedicated article.

Q1 2025

• Self-Service User Management: Enhances governance and control across your infrastructure.
  • Introduces role-based access: Viewer, Editor, Admin, Super Admin.
  • Define who can do what - fine-grained permissions via a permission matrix.
  • Assign access per ISA 95 structure, limiting users to specific sites or areas.
  • Simple onboarding: add name, email, assign role, define scope, and send invite.
  • Available now in the Management Console for all enterprise users

• Infrastructure-Level Audit Logs: Track every user action for full traceability
  • Logs every action on your infrastructure: login, data fetch, edits, user changes.
  • Works at the instance level, not just in the management console.
  • Exportable logs for validation - ideal for regulated environments like pharma.

• Git Diff for Deployment Transparency and Collaboration
  • Visual comparison between current and previous YAML versions.
  • Shows exactly what changed, who did it, and when.
  • Helps avoid deployment errors and improves team coordination.
  • Great for scaling teams working on complex infrastructure setups
• Made processing of data human again - “Processing Mode” Overhaul
  • ProcessingModeSelector: The feature flag for switching between Bloblang, Node-RED, Tag Processor, etc. is now enabled by default for all protocol converters. It was already announced in this blog post: https://learn.umh.app/course/node-red-within-benthos/
  • When selecting tag_processor as processing mode, we have implemented a nice UI that makes it even simpler to configure the protocol converter (see screenshots)

• New Intuitive OPC UA Input Configuration UI
  • Graphical OPC UA Setup: There’s a new dedicated UI for configuring OPC UA inputs—endpoint settings, advanced toggles, and auto-synced node IDs (see screenshot)
  • Extended Node ID Handling: We fixed scenarios where comments or unusual YAML structures would wipe node IDs. The UI better preserves them, even if you have large or commented YAML sections.

• Install UMH in Existing Kubernetes Cluster
  • User Choice: If you already have K3s running, the installer now asks if you want to skip re-installation. Handy for advanced setups where you just need to upgrade the Management Console or Helm charts.
    
• Making the User Interface even more intuitive
  • Enterprise License Display & Status
    • A new “License Status” section shows if you’re using an enterprise license.
    • If you already have an enterprise license, no upgrade prompts. If not, you’ll see an option to upgrade.
• Enhanced style fixes and clearer license validity dates.
• Tag Browser Smart Auto-Expand
  • Switched from width-based to height-based expansion.
  • Provides a better initial overview when browsing large sets of tags.
• Global App Filtering
  • A new “Filter” dropdown at the top helps you narrow data by location, instance, or other criteria.
  • If nothing matches, a “No results—clear filters?” prompt appears for quick resets.
• Fullscreen Code Editor
  • Edit large configs or code snippets without feeling cramped.
• Optional Bridge Headers (Kafka/MQTT)
  • Properly displayed when using Kafka.
  • Hidden when viewing MQTT to keep the interface protocol-specific.
    
• Making Benthos better for first-time & power users
  • Test Benthos Config (Beta/Feature Flag)
    • Directly test your Benthos input config in the Management Console.
    • A dedicated button (behind a feature flag) lets you quickly iterate by running test inputs.
  • Benthos Input Testing UI: Split-Pane Results
    • Collapsible entries show payload, metadata, and errors clearly.
    • No more cramped inline logs.
  • Benthos Image Tag Customization
    • Select a specific Benthos-UMH image tag in “expert mode” to test advanced or custom builds.
• Making OPC UA easier to use through browser enhancements
  • Extended Timeouts (10s → 300s) for slow responses.
  • Hierarchical Browsing toggle for servers with unusual responses.
  • Improved Visibility with extra styling/borders.
  • Retries if the root node is empty on the first try.

Other Improvements & Notable Changes

• Memory Usage & Performance
  • Reduced Companion Memory: We introduced deeper caching optimizations and pointer-based data structures for Kubernetes watchers. This drastically cuts memory overhead if you have many secrets, configmaps, or pods.
  • Retry & Backoff: Various fetch operations (companion version checks, instance polling, Helm re-checks) now automatically retry on transient network issues or DNS hiccups.
• Enhanced “DFC Rollback”
  • We remove resourceVersion in configmaps/secrets before attempting a rollback, making rollbacks more reliable and less prone to conflict errors.
• UMH Instances Polling
  • Split the instance data fetching vs. message processing into two separate steps, so the UI can show partial updates faster. Especially noticeable when you have many instances or slow networks.
• Auto-Scroll & Split Pane UI
  • Log Auto-Scroll: The log viewer only auto-scrolls if you’re at the bottom, so it won’t jerk back down if you’ve scrolled up.
  • Split Pane Scaling: Fixes for collapsed states and weird alignment issues, especially when toggling the left panel in code editors.
• DNS Handling / IPv4 Enforcement
  • The installer now checks if DNS recursion is available over IPv4. If not, it prompts you for alternative DNS servers, preventing certain “DNS misbehaving” errors on hosts like Hetzner.
• Helm & Installer
  • Limit Helm Revisions: We added MaxHistory=5 and CleanupOnFail so the cluster doesn’t accumulate hundreds of old Helm secrets.
  • Atomic Installs: Ensures that a Helm upgrade either completes fully or is rolled back, reducing partial-install headaches.
• Go Version Bump
  • Updated internal build environment from Go 1.23.x → 1.24.x, offering better performance and some updated language features.
• Performance & Stability
  • Raised default QPS & Burst for high-throughput (50k tags/sec!)
  • Better concurrency handling and faster simultaneous downloads.
• SQL Query Resolution Selector
  • In the Tag Browser’s SQL tab, choose different time resolutions or raw data.
• ActionErrorCollapsible for DFCs
  • A top banner shows consolidated error messages for easier troubleshooting.
• Duplicate Deployment Checks for DFCs
  • Prevents multiple DFC deployments by the same connection ID.
• License Box Style
  • Cleaned up styling on the homepage for consistency.
• Svelte Dependencies
  • Removed svelte-chartjs in favor of direct Chart.js for a smaller bundle.
• Refined Applications & Instances
  • Improved IP display (real names + statuses).
  • Instances show Management Companion version (not just Helm chart version).
  • Added fallback sorting by UUID for stable listings.
• Auto-Scrolling Log Boxes
  • Always see the latest log messages without manual scrolling.
• Rocky Linux Auto-Update
  • Enforces certain minor versions and auto-updates for better security.

Q4 2024

• User Experience Enhancements
  • Making SQL Queries Even Easier to Build with the Resolution Selector: The Tag Browser now includes a resolution selector that shows available options like $__interval, 1 hour, 1 minute, and 1 second, improving query flexibility.
  • Sample library to push data from the UNS to other applications : A protocol dropdown list with a wide array of data sinks and sources has been added when creating custom data flow components to simplify the configuration.

• System and Performance Improvements
  • Enabling the companion to check the disk health: The Management Companion now verifies instance stability and disk availability before upgrading to prevent potential failures.
  • Experience the UMH is now even easier with more sample data: We expanded the Demo Simulator to cover more verticals, addressing a wider range of use cases and sample industries.
    
• Multitask with multiple tabs with the Management Console
  • We understand that there are multiple workflows where you want to "model your UNS data"; therefore, we added the ability to open multiple tabs simultaneously.
  • This gives the user the ability to more easily copy and paste between processing steps and also monitor the data going through the pipeline, among other things.
  • As a nice side effect, the management console also got way faster due to the underlying technical changes

• Better reliability and parsing in benthos-umh
  • Fixed a bug in sensorconnect that resulted in "hanging up" after some time (thank you, @EmlingerM!)
  • Fixed a bug with sensorconnect where octetStrings were not properly parsed
  • Cleaned up the module imports (chore)
• Management Companion updates
  • Fixed a bug where on some systems k3s would fail to start due to misconfigured fs.inotify.max_user_watches and fs.inotify.max_user_instances
  • Added stricter instance name validation, preventing invalid k3s node names
  • Added automatic recovery for stuck Helm upgrades by attempting to rollback to the new version after 5 minutes
  • Offloaded some of the processing in the frontend to a dedicated worker, resulting in better reactivity
  • Fixed a bug where upgrading from old Helm charts would sometimes fail
  • Added more machines to our demo simulator
  • Added release channels and dedicated enterprise channels
• Streamline Data Discovery on OPC UA Servers with Our OPC UA Browser
  • We have added to our Protocol Converter feature in the management console the ability to browse and explore the configured OPC UA Server
  • This makes it much easier to actually see which data points are available and which are not, thereby reducing the time to value per new OPC UA Server
  • You can select a set of folders or individual nodes and apply those directly to your input YAML to streamline the workflow even further.

• Connect IO-Link with the UNS and UMH
  • You can now use IO-Link sensors to enrich the UNS with tactical process and production data like temperature sensors, light barriers, or vibration data through the plugin in the protocol converter
  • This reduces time spent on additional middleware by connecting directly to the sensor and gateway
  • It automatically parses the hex-encoded sensor data from the IO-Link Master with the corresponding IODD files to create a human-readable JSON with all values in key-value format
  • All data points are then available in the tag browser and are automatically persisted in the database

• New Benthos-based bridges and fully Automated Helm Upgrade from the Management Console
  • We have now ported all our internal bridges (e.g., MQTT -> Kafka, Kafka -> MQTT) to Benthos-based DFCs to increase performance, maintainability, and directly make features like logs and metrics available in the management console
  • The Upgrade button has now become even more powerful! Now, not only does the management companion get updated from the management console, but also the attached Helm chart of the UMH, making upgrading and staying up to date even easier

Q3 2024

• New Homepage and small UI rework
  • We have added a new homepage that now displays relevant usage KPIs such as tags captured, current throughput, and connected interfaces. It also includes a list of alerts for your entire deployment that require your attention.
  • Additionally, a list of the latest UMH blog posts is now available to help you learn about or improve your deployment.
  • Help pages are now accessible via a side panel that you can slide in on relevant pages to assist you in navigating quickly.

• Templates for DFCs to help you bridge data more easily
  • Introduced new custom Data Flow Control (DFC) templates for kafka-to-postgres and other bridges. This preview feature allows you to store your own data in the database or bridge your custom data contracts between MQTT and Kafka.
• Easily Upgrade Your OS via the Management Console
  • A clear upgrade path for Flatcar and Rocky Linux has been implemented. Users will now receive detailed messages explaining why a version is outdated, along with guidance and an easy command to update.
• New Instance View and Tab feature: The new Instance view replaces the old Component View and increases amount of visible information and improves the ability for non technical users to use & troubleshoot the infrastructure.

• Data Flow Components: We’ve enabled the deployment of fully custom Benthos setups to make all Benthos use cases available through the Management Console. We also included a template, e.g., mirroring the UNS to AWS SQS, - more to follow soon.
• Released Modbus over TCP/IP Connector for Benthos-UMH: Added the ability to connect to devices via Modbus over TCP/IP e.g. energy meters, which is fully end-to-end tested against a Modbus simulator. Additional documentation is available here.
• Integrated Protocol Converter into Data Flow Components: Protocol converters are now deployed using the same mechanism as custom data flow components, improving reliability and enabling version control and rollbacks.
• ARM64 Support for UMH Lite: Initial support for ARM64 is available for UMH Lite deployments, increasing accessibility for ARM-powered devices for data capture. UMH Classic will follow at a later time.
• Included CyberChef / Quick data conversion: We’ve integrated the popular tool "CyberChef" for simple and complex data manipulation, such as decoding Base64 or Hex payloads. It's available for every data point in the tag browser.