2026-09-18

Marker option category filtering now works consistently across legacy and current data formats, with category information included in responses

Hub API

Marker option category filtering now works consistently across legacy and current data formats, with category information included in responses
Improved media region filtering performance and reliability by using optimized geospatial query paths
Media responses now include source vault and source media identifiers to better support on-demand recording workflows
Case media listings now automatically restore missing FPS metadata from source media, improving playback and media handling accuracy

Added optional owner-only developer tools on device pages, including live Agent logs and terminal access
Improved media playback reliability and compatibility with upgraded streaming and video player support
Enhanced media browsing with configurable priority ordering for marker filter categories
Expanded media player capabilities with improved playlist handling, seek controls, and download support

Added optional debug-level processing diagnostics with traceable checkpoint logs to help investigate pipeline event handling issues
Improved processing reliability by detecting and failing database updates that do not modify the expected activity or latest-media records
On-demand forwarding now preserves source ownership and source media metadata during processing
Forwarded persist events now retain their original completion timestamp through monitor enrichment
Encrypted payload metadata is now preserved correctly when pipeline events are processed and forwarded

Improved calendar and available-day processing to immediately reflect organisation ownership, timezone, and retention changes without requiring cached data refreshes
Enhanced reliability for concurrent event ingestion, reducing the risk of duplicate or conflicting available-day updates during high-volume processing
Optimized date handling performance for large or unsorted recording histories while preserving chronological retention behavior for late-arriving recordings
Added bounded retry handling for database update contention to improve stability under load and prevent runaway retry loops
Introduced an optional forced dead-letter replay mode for safely testing queue recovery and replay workflows in isolated environments
Expanded structured logging and distributed tracing support to improve visibility into processing outcomes, retries, and dead-letter routing across services
Added validation for incoming tracing context to improve cross-service request correlation and diagnostics
Improved timeout handling so ingestion tasks respect caller deadlines and cancel cleanly when requests are interrupted

Improved reliability when processing media with expired or rejected signed URLs by automatically refreshing access links and retrying downloads
Added retry handling for temporary media download failures and transient server errors to reduce interrupted processing jobs
Enhanced compatibility with custom and overridden storage providers when retrieving media files

Improved thumbnail generation reliability and image quality by updating frame extraction and JPEG encoding behavior
Thumbnail uploads now use JPEG format consistently, improving compatibility with downstream consumers
Added automatic recovery for expired or rejected media signed URLs, reducing failed thumbnail processing jobs
Added support for Vault storage override credentials and providers when downloading source media

Improved reliability when processing media files by automatically refreshing expired or rejected signed URLs
Reduced pipeline failures caused by expired storage links during media downloads
Improved handling of temporary download errors with more resilient retry behavior for media retrieval

Added support for managing Vault storage providers, integrations, and accounts through environment-based configuration with validation and stable deployment identities
Environment-defined accounts, integrations, and storage providers are now protected as read-only to prevent accidental modification or deletion
Added new Outbox monitoring APIs to view delivery backlogs, recent outcomes, throughput, failures, and latency for integrations
Improved device deletion reliability by using stable camera keys instead of database object IDs
Simplified and updated MQTT and forwarding configuration options, including new concurrency controls and deployment identity settings
Added support for persistent Vault deployment IDs to improve consistency across replicas and restarts

Improved documentation navigation and cross-linking across Agent, Factory, Vault, Hub, and Enterprise guides for a smoother browsing experience
Added a new option to automatically apply the current user as the default assignee filter in Hub case lists
Fixed multiple broken and outdated documentation links related to analytics, recordings, integrations, storage, and deployment setup

Added a new audit service with configurable webhook destinations for forwarding and integrating audit events
Improved audit event reliability with durable intake queues, retry handling, dead-letter processing, and lease-based delivery coordination
Added frontend customization for shared chart and detection overlay colors
Added feature flags to enable or disable device analytics, developer tools, and fullscreen dashboard site maps
Expanded deployment configuration options for managing audit service scaling, polling intervals, retention behavior, and custom environment settings