Member Junction
    Preparing search index...

    Interface TestRunResult

    Result from running a single test

    interface TestRunResult {
        completedAt: Date;
        durationMs: number;
        errorMessage?: string;
        failedChecks: number;
        oracleResults: OracleResult[];
        passedChecks: number;
        replay?: ReplayTelemetry;
        resolvedVariables?: ResolvedTestVariables;
        score: number;
        sequence?: number;
        skippedChecks?: number;
        startedAt: Date;
        status: "Passed" | "Failed" | "Skipped" | "Error" | "Timeout";
        targetLogEntityId?: string;
        targetLogId: string;
        targetType: string;
        testId: string;
        testName: string;
        testRunId: string;
        tier?: string;
        totalChecks: number;
        totalCost: number;
    }
    Index

    Properties

    completedAt durationMs errorMessage? failedChecks oracleResults passedChecks replay? resolvedVariables? score sequence? skippedChecks? startedAt status targetLogEntityId? targetLogId targetType testId testName testRunId tier? totalChecks totalCost

    Properties

    completedAt: Date

    When execution completed

    durationMs: number

    Execution duration in milliseconds

    errorMessage?: string

    Error message if status is Error

    failedChecks: number

    Number of checks that failed

    oracleResults: OracleResult[]

    Oracle evaluation results

    passedChecks: number

    Number of checks that passed

    replay?: ReplayTelemetry

    Replay telemetry, present whenever a replay was ATTEMPTED — so the drift signal (diverged > 0) survives even a green LLM-fallback result. Absent on pure-LLM runs. See ReplayTelemetry.

    resolvedVariables?: ResolvedTestVariables

    Resolved variables that were used for this test run

    score: number

    Overall score (0.0000 to 1.0000)

    sequence?: number

    Iteration number for repeated tests (when RepeatCount > 1)

    skippedChecks?: number

    Number of checks skipped (tier-gated / environment gaps) — counted separately from the executed totals so a partially-skipped run is visibly not full coverage.

    startedAt: Date

    When execution started

    status: "Passed" | "Failed" | "Skipped" | "Error" | "Timeout"

    Test execution status

    targetLogEntityId?: string

    Entity ID identifying the type of target being tested. References Entity.ID (e.g., Entity ID for "MJ: AI Agent Runs").

    targetLogId: string

    Target entity ID (e.g., AIAgentRun.ID)

    targetType: string

    Optional sub-category or variant label for the test target. Use for ad-hoc labeling or to distinguish test scenarios within the same entity type. Examples: "Summarization", "Classification", "Code Review", "Multi-turn Chat"

    testId: string

    Test ID

    testName: string

    Test name (from lookup field)

    testRunId: string

    Test Run ID

    tier?: string

    Execution-tier label a tiered driver reports — e.g. Computer Use's 'replay' / 'replay-with-heal' / 'llm'. The tier that PRODUCED this result: a replay that diverged and fell back reports 'llm'. Reporting segments tier mix / replay share by it. Absent for single-tier drivers.

    totalChecks: number

    Total number of checks executed (passed + failed)

    totalCost: number

    Cost in USD

    Settings

    Member Visibility

    On This Page

    Properties