Configuration

VrsHealthCheck has an important feature called check configurations. Each check configuration defines a collection of check criteria and thresholds. This feature allows a single, unified VrsHealthCheck binary to handle all divergent checks, simply by taking different configurations.

Example configurations:

1. Default Configuration: A baseline health check across all VRS streams, focused on data completeness and detecting hardware, firmware, or OS-level failures.
2. Algorithm-Specific Configurations: For each algorithm workflow, such as location, hand tracking, or eye tracking, a distinct set of checks can be created to assess the suitability of the VRS for running the algorithm. For example, the location configuration might impose stricter requirements on SLAM and IMU sensors.

Listing Available Configurations

To see all available threshold configurations:

run_vrs_health_check --list-configurations

Example output:

Available VrsHealthCheck configurations: AriaGen2_Default, AriaGen2_Location, AriaGen1_Default, AriaGen1_Location

Viewing Configuration Details

To see the detailed checks and thresholds for a specific configuration:

run_vrs_health_check --show-configuration-json AriaGen2_Default

This displays the complete JSON structure with all threshold values for that configuration.

Selecting Configuration for Output

To control which configuration's results are reported as the final status, use the --choose-configuration flag:

run_vrs_health_check --path recording.vrs \
  --output-json results.json \
  --choose-configuration AriaGen2_Location

⚠️ Important Note: The --choose-configuration flag does not limit which configuration's checks are performed. All available configurations are evaluated during every run and the results are reported in the output.json file. This flag only controls:

1. Final Exit Code: Which configuration's result determines the tool's exit status (PASS/WARN/FAIL)
2. Terminal Output: Which configuration's detailed results are displayed in the console

This design ensures you get comprehensive analysis data for all configurations while allowing you to focus on the specific threshold set most relevant to your use case.

Understanding Checks

Checks are quality validation rules applied to VRS data streams. Each check evaluates a specific metric against predefined thresholds. There are two fundamental types of checks: Numeric Checks and Boolean Checks.

1. Numeric Checks

Numeric checks compare measured values (like frame rates, error counts, time intervals) against numerical thresholds.

Supported Check Types:

• LE (Less than or Equal): Pass if measured_value <= threshold
• LT (Less than): Pass if measured_value < threshold
• GE (Greater than or Equal): Pass if measured_value >= threshold
• GT (Greater than): Pass if measured_value > threshold

Structure:

{
  "check_type": "LE|LT|GE|GT",
  "fail_threshold": <number>,     // Optional: value that triggers FAIL
  "warn_threshold": <number>      // Optional: value that triggers WARN
}

Evaluation Logic:

1. PASS: Value meets all thresholds
2. WARN: Value exceeds warn_threshold but not fail_threshold
3. FAIL: Value exceeds fail_threshold

2. Boolean Checks

Boolean checks verify true/false conditions like calibration validity or file integrity.

Supported Check Types:

• EQ_TRUE: Pass if value is true
• EQ_FALSE: Pass if value is false

Structure:

{
  "check_type": "EQ_TRUE|EQ_FALSE",
  "fail_if_mismatch": true|false    // Whether mismatch causes FAIL or WARN
}

Evaluation Logic:

• PASS: Value matches expected boolean
• FAIL/WARN: Value doesn't match (severity depends on fail_if_mismatch)

---