WEDA Core v1.2.0

Release date: October 29, 2026

WEDA Core v1.2.0 expands remote device operations, external device integration and data processing. It also adds WEDA Console workflows and changes API fields, validation and deployment-result reporting. The changes below are relative to v1.1.2.

What's New

Area	Added in v1.2.0
Web SSH	Open a browser-based SSH terminal through a WEDA Tunnel session. Integrate the returned Web SSH URL into your application and sign in with the target device's SSH account. The terminal includes file upload and download.
Device heartbeat	Read connectivity.state and connectivity.updatedAt through the device APIs. States distinguish Connected, Abnormal, Disconnected and Unknown. WEDA Nodes and SubNodes have their own connectivity state.
Remote commands	Request reboot, shutdown or a command declared by the target device, then track its result using reqSeqId. Supported Advantech devices can expose GPIO digital-input and digital-output commands when enabled in their device configuration.
Data Resync	Request re-upload of locally recorded telemetry for a selected past time range on a device that supports report.historical.
Telemetry configuration	Discover supported sensors and configuration schemas, configure reporting and local recording separately, inspect apply results and reset desired settings.
Custom Devices and Client Credentials	Integrate external devices and applications through HTTP or MQTT using a Data Source and Data Parser. HTTP ingestion uses Client Credentials for non-interactive authentication. Input supports values, structured Data Packs and blob files, with APIs for reading Data Packs and blob metadata.
Data aggregation	Query aggregates over a selected time range using ten supported functions, including AVG, SUM, COUNT, FIRST_VALUE and LAST_VALUE. Aggregation is performed when queried.
JavaScript and Multilayer Analyzers	Process up to ten input DataStreams and publish multiple outputs. Connect a JavaScript Analyzer's output DataStream to another JavaScript Analyzer, and configure how multiple inputs are assembled for processing.
MQTT Data Bridge TLS	Forward DataStream values to MQTT brokers using mqtts://. Manage uploaded CA certificates separately and reference them in a broker configuration.
WEDA Console	Manage AI Models and Editions, pause or resume model uploads, deploy to selected devices and track deployments. New DataStreams, Analyzers, DataStream Bridges and MQTT Broker Configs views support data-management workflows.

WEDA Console source is available for customers and partners building custom applications. WEDA Console and WEDA Core versions are managed separately; confirm compatibility for your environment.

Behavior and API Changes

For endpoint-by-endpoint changes, see API Changes in v1.2.0.

Caution: Review integrations before upgrading

This release changes response fields, accepted request values and error handling. Update clients that depend on the v1.1.2 contract before moving them to v1.2.0.

Area	Change and required action
Device lists	Organization device lists can include WEDA Nodes, SubNodes and Custom Devices. Branch on type, handle a nullable hwModel, use createdAt for creation-time sorting and keep maxResultCount at or below 100.
Device registration and updates	Device identifiers, hardware models and names have stricter validation. Device names must be unique within an Org, ignoring case. Device updates reject changes to deviceId, hwModel and orgId; use the supported Org device-linking operation for association changes.
Response timestamps	Affected Device, Tunnel, Container and Organization JSON responses replace legacy keys such as creationTime, lastModificationTime, dateCreated and dateModified with createdAt and updatedAt. The old keys are no longer returned. Update clients that read those keys.
Telemetry settings	Prefer report for sensor reporting settings; the legacy config key is still accepted, but both cannot be sent for the same sensor. Supply enabled and interval explicitly. An enabled record setting also requires its recording interval. GET prefers desired settings and falls back to reported settings when no desired configuration exists; use apply-result fields to check device application.
AI Model deployment results	DOWNLOADED is replaced by DOWNLOADING, VERIFYING and STAGING. Update state enums and filters. rejectedReason identifies a failure category; use errorCode and errorMessage for device error details.
AI Model file download	The model-edition file-download API (GET / HEAD) is removed. Remove download controls and integrations that depend on this API.
Container command results	Results include the command phase and structured error details. Device-reported executedAt and completedAt can be null until received. Distinguish command acceptance from the device's execution outcome.
Registry configuration	Registry hosts must be unique within an Org. A conflicting create or host change returns 409 RegistryHostExists; review configurations that use the same host with different credentials.
MQTT broker updates	Duplicate broker configuration names return 409. A PUT without tls clears the uploaded CA reference; retain and resend tls when preserving that configuration.
Errors and permissions	Use error title for classification and validationItems for field validation details. Missing-device responses change from 403 to 404 in Tunnel and AI Model Management. Tunnel creation requires an Org Admin, and terminal-state conflicts return 409.

Fixed

• Registry-name uniqueness checks on update now apply within the Org instead of the Tenant, avoiding conflicts with names used in other organizations.
• Invalid stack-deployment parameters and invalid deployment sorting return 400 Bad Request instead of 500 Internal Server Error.

Limitations

• Data Resync: Recovery requires local recordings that are still retained on the device, within a seven-day horizon. Replayed values overwrite cloud values with the same timestamp. Recovery must be requested; it is not automatic after every disconnection.
• JavaScript Analyzers: Scripts are limited to 10,000 characters and ten input DataStreams. Chains support JavaScript Analyzers only, with a maximum depth of five and no cycles. Each evaluation has configured execution budgets; defaults are 2,000 ms, 4,000,000 allocated bytes and 10,000 statements. Exceeding a budget produces no output for that evaluation.

Upgrade and Compatibility Notes

1. Match the Core and Node major.minor versions. Use WEDA Node v1.2.x with WEDA Core v1.2.x for full v1.2 feature support. Patch and build numbers can differ. Check the target SubNode SDK and declared capabilities for device-dependent operations.
2. Review existing device records and client mappings. Check identifiers and names against the new rules, and update response timestamp names, deployment-state handling, nullable command timestamps and error parsing.
3. Review saved configurations. Send explicit telemetry reporting fields and complete nested objects when replacing them. Preserve the TLS reference when updating a broker configuration, and check registry-host conflicts within each Org.
4. Verify results after the upgrade. Check device heartbeat status, configuration apply results and device-reported deployment or command outcomes. An accepted request alone does not confirm device execution or application readiness.

---

Last updated on Oct-27, 2026 | Version 1.2.0