New form of analytics
All original visualizations based on D3 have been replaced by more than twenty new visualizations created using ECharts. Each chart works with current data and allows export to CSV format.
Organizers have access to a new dedicated dashboard for training analysis, which offers three views:
- Aggregated view — provides an overview of the entire training instance: status, progress over time, scoreboard, number of players at individual levels, score attainment, level difficulty, most frequent incorrect answers, and analysis of commands used.
- Trainee — allows one specific training run to be followed from beginning to end: cumulative score, relationship between time and achieved score, event timeline, and a complete command log including the host, source IP address, and working directory.
- Assessment — displays the distribution of answers for individual questions and thus supports the analysis of question quality and difficulty. Data can also be exported to CSV.
Trainees will now receive a summary of their training run, showing how their attempt ended and how they performed within the given instance compared with others.
Performance
All charts are powered by a new event processing system. Events are synchronized into a local database directly in the browser, so charts work with local data instead of repeatedly loading data over the network.
For each instance and event type, the system tracks which events have already been loaded, so only new records are downloaded. Updates are also paused when the user does not have the given browser tab open.
Redesigned score export
Scores are now divided by individual levels, and the assessment score is separated from the training score.
The user interface allows users to export a CSV file containing rank, participant identity, state, timing information, scores for individual levels, hints used, incorrect answers, and displayed solutions.
The API now provides a corresponding structured report in JSON format.
Sandbox and infrastructure
- Support for network traffic mirroring (port mirroring) for OpenStack (Neutron TaaS with OVN) and AWS (VPC Traffic Mirroring on Nitro instances).
- Configurable caching for the GitHub sandbox-definition provider.
- Event storage has been migrated from Elasticsearch to OpenSearch.
- Windows sandbox hosts now use PowerShell by default for SSH communication.
- Migration to Python 3.12+ and removal of remaining known security vulnerabilities (CVEs) in the sandbox and Guacamole services.
Retired components
Adaptive Training, Adaptive Smart Assistant, and the Elasticsearch service have been marked as deprecated and archived.
The MITRE techniques service now provides a single unified endpoint for training techniques.
📌 Complete release notes on GitHub