fix(exams): add Sentry capture + better logging for AI grading errors

Grading errors were silently swallowed in try/catch. Now logs the
grading result error and reports exceptions to Sentry with examId
and submissionId context.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: guillermoscript

app/[locale]/dashboard/student/courses/[courseId]/exams/[examId]/exam-taker.tsx

@@ -139,11 +139,14 @@ export function ExamTaker({
       // Trigger AI grading
       try {
         const { gradeExamWithAI } = await import('@/app/actions/exam-grading')
-        await gradeExamWithAI({
+        const gradingResult = await gradeExamWithAI({
           examId,
           submissionId: submission.submission_id,
           answers,
         })
+        if (!gradingResult.success) {
+          console.error('AI grading returned error:', gradingResult.error)
+        }
       } catch (gradingError) {
         console.error('AI grading failed (non-blocking):', gradingError)
         // Continue to results page even if AI grading fails

app/actions/exam-grading.ts

@@ -454,6 +454,11 @@ Evaluate these free-text answers now:`
     }
   } catch (error) {
     console.error('Error grading exam:', error)
+    // Report to Sentry for observability
+    try {
+      const Sentry = await import('@sentry/nextjs')
+      Sentry.captureException(error, { extra: { examId: params.examId, submissionId: params.submissionId } })
+    } catch {}
     return {
       success: false,
       error: error instanceof Error ? error.message : 'Failed to grade exam',