Skip to main content
Glama

dashboard

Monitor RLHF performance metrics including approval rates, gate statistics, prevention impact, and system health status for comprehensive oversight.

Instructions

Get full RLHF dashboard -- approval rate, gate stats, prevention impact, system health

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault

No arguments

Implementation Reference

  • This function generates the data object for the "dashboard" tool.
    function generateDashboard(feedbackDir, options = {}) {
      const analyticsWindow = resolveAnalyticsWindow(options.analyticsWindow || options);
      const feedbackLogPath = path.join(feedbackDir, 'feedback-log.jsonl');
      const diagnosticLogPath = path.join(feedbackDir, 'diagnostic-log.jsonl');
      const entries = readJSONL(feedbackLogPath);
      const diagnosticEntries = readJSONL(diagnosticLogPath);
      const billingSummary = options.billingSummary || getBillingSummary(analyticsWindow);
    
      const approval = computeApprovalStats(entries);
      const gateStats = computeGateStats();
      const prevention = computePreventionImpact(feedbackDir, gateStats);
      const trend = computeSessionTrend(entries, 10);
      const health = computeSystemHealth(feedbackDir, gateStats);
      const diagnostics = aggregateFailureDiagnostics([...entries, ...diagnosticEntries]);
      const secretGuard = computeSecretGuardStats(diagnosticEntries);
      const analytics = computeAnalyticsSummary(feedbackDir, {
        analyticsWindow,
        billingSummary,
      });
      const observability = computeObservabilityStats(diagnosticEntries, diagnostics, secretGuard, analytics.telemetry);
      const instrumentation = computeInstrumentationReadiness(analytics, billingSummary);
      const delegation = summarizeDelegation(feedbackDir);
      const readiness = generateAgentReadinessReport({ projectRoot: PROJECT_ROOT });
    
      return {
        operational: {
          source: options.billingSource || 'local',
          fallbackReason: options.billingFallbackReason || null,
          window: analytics.window || analyticsWindow,
        },
        approval,
        gateStats,
        prevention,
        trend,
        health,
        diagnostics,
        delegation,
        secretGuard,
        analytics,
        observability,
        instrumentation,
        readiness,
      };
    }
  • Tool handler for the "dashboard" tool.
    case 'dashboard':
      return toTextResult(generateDashboard(getFeedbackPaths().FEEDBACK_DIR));

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/IgorGanapolsky/mcp-memory-gateway'

If you have feedback or need assistance with the MCP directory API, please join our Discord server