OpenFrame v0.7.8 - Platform Architecture & Monitoring

Comprehensive documentation ecosystem covering API references, deployment guides, user manuals, integration guides, and architectural documentation. Includes versioned documentation with installation procedures, configuration options, API endpoints with examples, troubleshooting guides, best practices, and security considerations.

Integrated FleetMDMM as an advanced remote monitoring and management solution providing real-time device monitoring, automated alerting, patch management, remote access capabilities, and comprehensive reporting dashboards. Enables MSPs to monitor endpoint health, automate maintenance tasks, deploy software updates, and maintain complete visibility across managed devices.

Split the existing gateway into two dedicated gateways: User Gateway for end-user client requests and Agent Gateway for agent requests, improving security isolation and performance.

Real-time Fleet dashboard updates via WebSocket connections for live device status and monitoring data.

Enhanced remote control capabilities with full clipboard synchronization between local and remote machines during MeshCentral sessions.

Added support for scheduling and running scripts on macOS devices with one-time, recurring, and cron scheduling options.

Comprehensive policy management and query capabilities for Fleet device management.

Real-time status indicators for all tool agents (Tactical, Fleet, osquery) showing online/offline status and last seen timestamps in the Agent tab.

Fixed issue where dashboard data refetched incorrectly or excessively, causing unnecessary API calls and UI flickering.

Resolved inconsistent pagination results when using status sorting, where devices appeared duplicated across pages and counts didn't match database records.

Eliminated flickering in navigation bar and centered icons properly in minimalist version.

Fixed logout flow to consistently clear session and redirect to login page immediately without requiring manual refresh.

Corrected device view to display actual OpenFrame last seen timestamp instead of last boot time from TacticalRMM.

Uninstall process now automatically stops and terminates MeshCentral service, eliminating need for manual intervention.

Fixed database error that occurred when combining pagination with search queries.

Resolved timeout errors when processing approval requests by implementing asynchronous processing for long-running commands.

Fixed remote control sessions to properly capture Shift key, enabling special characters and capital letters to pass through correctly.

Resolved issue where Fae application would pop up unexpectedly during server offline periods or self-updates.

Restored Fleet agent connectivity following migration to fork repository.

Fixed SQL error during Fleet migration process when adding policy_id column that already exists.

Resolved port configuration issues in CI/CD pipeline causing test failures in self-hosted environments.

Fixed abnormally high number of socket connections to Tactical RMM that caused gateway performance degradation.

Reduced MongoDB connections from ~1,800 to 180-360 across cluster through connection pool tuning and removal of duplicate MongoClients.

Improved skeleton placeholders on Device and Script pages to better match actual page layout for smoother loading experience.

Consolidated Logs Details component implementations into single reusable component for consistent behavior across all pages.

Replaced client-side script loading with dynamic server-side autocomplete to handle large script libraries without performance issues.

Improved NATS WebSocket reliability with proper reconnection logic, chunk catchup, token refresh, and subscription cleanup to prevent memory leaks.

Enabled self-update functionality for MeshCentral and Fae applications running as system services on Mac and Windows.

Updated MeshAgent permission window to follow OpenFrame design standards on macOS.

Split monolithic core-data library into separate modules per data system (Cassandra, NATS) to reduce classpath bloat and improve startup times.

Implemented exponential backoff for agent connection retries to reduce server load during outages.

Added automatic time zone detection in TacticalRMM agents, eliminating manual configuration requirement.

Added scheduled job to detect and mark offline devices when NATS events are unavailable, running every 60 seconds with distributed locking.

Enhanced Scripts v2 API with advanced filtering, search capabilities, and cursor-based pagination for scalable performance.

Reduced redundant authentication checks and improved refresh token handling for better session stability and performance.

Implemented consistent, structured error responses across all auth endpoints with meaningful user feedback.

Created shared Tailwind configuration in unified library ensuring consistent styling, spacing, colors, and design tokens across all projects.

Moved HubSpot subscription, verification emails, and other post-registration processes to async pipeline, eliminating 504 timeout errors.

Improved KeyGenerator with NativePRNGNonBlocking provider, configurable key size and algorithm, and automatic Key ID generation.

Moved EncryptionService to dedicated module, removing unnecessary spring-security-crypto dependency from services that don't use encryption.

Created dedicated shared library for common error handling logic including global exception handlers, base exception types, and standardized error responses.

Extended Storybook configuration to include shared component library for unified documentation and testing.

Increased NATS retry strategy with higher retry count and backoff to ensure reliable subscription recovery during unstable connections.

Moved verbose tool agent logs (token refresh, SSL certificate messages) to debug level to prevent log pollution.

Made actual agent versions configurable through API for flexible version management.

Implemented Redis-based locking during chat processing to prevent parallel execution conflicts.

Enabled LLM to determine appropriate command execution timeout instead of fixed 30-second limit.

Added exponential backoff for NATS WebSocket calls to prevent server overload during reconnection attempts.

Enhanced Mac remote connection reliability with improved connection handling, reconnection logic, and error recovery mechanisms.

Migrated Fleet Dockerfiles and container images to dedicated forked repository for independent versioning and faster CI/CD.

Implemented pre-check validation for Fleet migrations to prevent repeated executions and ensure safe reruns without breaking cluster state.

Added clickable device links in test reports for direct navigation to device details pages.

Moved legacy agents into separate group for improved clarity and isolation from active agents.

Pending devices now hidden by default in device list view, showing only active devices unless explicitly included.

Added user-friendly error page with clear messaging and retry guidance when tenant gateway is unreachable.

Set up Debezium CDC connector for Tactical PostgreSQL autotask table with field mapping and error handling.

Added ZooKeeper PROPERTYSTORE cleanup to Pinot table deletion flow for complete resource cleanup.