Granola Incident Runbook

Granola Incident Runbook

Overview

Standard operating procedures for Granola incidents — from individual recording failures to organization-wide outages. Covers triage, remediation, communication, escalation, and post-incident review. Designed for IT admins, team leads, and individual users.

Prerequisites

• Granola admin access (for org-level incidents)
• Bookmark status.granola.ai for service status
• Internal communication channel identified (#granola-support or similar)

Instructions

Step 1 — Triage: Assess Severity

# Quick status check
curl -s "https://status.granola.ai/api/v2/status.json" 2>/dev/null | python3 -c "
import json, sys
try:
    data = json.load(sys.stdin)
    indicator = data.get('status', {}).get('indicator', 'unknown')
    desc = data.get('status', {}).get('description', 'Unknown')
    print(f'Status: {indicator} — {desc}')
except:
    print('Cannot reach status page — possible network issue')
" || echo "Network error — check internet connection"

Scope assessment questions:

1. Is it just you, or are others affected too? → Ask in #granola-support
2. Is status.granola.ai showing an incident? → P1/P2 if yes
3. Was it working earlier today? → Recent change (OS update, permissions) likely
4. Which platform? (Zoom/Meet/Teams) → Platform-specific audio routing

Step 2 — Remediation by Incident Type

Incident: "Meeting Not Recording"

Severity: P3 (single user) or P2 (team-wide)

Immediate actions:

1. Click Granola menu bar icon > Start Recording (manual override)
2. If manual start fails:
   • Check that the meeting has audio playing
   • Verify Granola is running: pgrep -l Granola
   • Restart: right-click menu bar icon > Restart Granola

Root cause investigation:

• Calendar event has a video conferencing link (Zoom/Meet/Teams)
• Calendar is connected in Settings > Calendar
• Microphone permission granted
• Screen & System Audio Recording permission granted (macOS)
• Not running conflicting audio software (Loopback, BlackHole)

Backup: Take manual notes. After fixing, re-record the next meeting.

Incident: "Transcription Stops Mid-Meeting"

Severity: P3

Immediate actions:

1. Check: is the computer awake? (Sleep kills transcription)
2. Right-click Granola icon > Restart Granola
3. Reopen your note — transcription may resume
4. If using Bluetooth: switch to wired audio or built-in speakers

Root cause: Granola stops transcription after ~15 minutes of no detected audio. Bluetooth devices can cause intermittent dropouts.

Incident: "Enhancement/Processing Failed"

Severity: P3

Immediate actions:

1. Wait 15 minutes — long meetings take longer to process
2. Check internet connectivity
3. Check status.granola.ai for service issues
4. Restart Granola and reopen the note
5. Click Enhance Notes again

If still failing after 30 minutes: The transcript was captured but enhancement may be queued. Submit support ticket with the meeting date/time.

Incident: "Integration Not Syncing" (Slack/Notion/HubSpot)

Severity: P3

Immediate actions:

1. Settings > Integrations > check the target integration status
2. Disconnect and reconnect the integration
3. Re-share the note manually
4. For Zapier: check Zap history at zapier.com for errors

Common causes:

Incident: "Granola Service Outage" (P1)

Severity: P1 Critical

Immediate actions:

1. Confirm at status.granola.ai
2. Switch to backup note-taking immediately:
   • Open a text editor or Google Doc
   • Take manual notes for active meetings
   • Notes can be combined with future Granola captures manually
3. Communicate to your team via Slack:

:rotating_light: Granola is currently experiencing a service outage.

Status: https://status.granola.ai
Impact: Meeting recordings and AI enhancement are unavailable.

Workaround: Take notes manually in Google Docs or your preferred editor.
I'll update when service is restored.

Next update: [time + 30 min]

1. Subscribe to status updates at status.granola.ai
2. Resume normal operation when status returns to Operational

Step 3 — Escalation Path

Level 1: User Self-Service
  → Restart Granola, check permissions, verify audio
  → Time: 5 minutes

Level 2: IT Support / Team Admin
  → Run debug bundle (granola-debug-bundle)
  → Check org-wide status, verify SSO/SCIM
  → Time: 15-30 minutes

Level 3: Granola Support
  → Submit ticket at help.granola.ai
  → Attach debug bundle
  → Enterprise: Priority support, dedicated contact
  → Time: 1-24 hours depending on severity

Level 4: Granola Engineering (P1 only)
  → Escalated by Granola Support for service outages
  → Status page updates provided by Granola team

Step 4 — Post-Incident Review

After resolution, document:

## Post-Incident Report

**Date:** YYYY-MM-DD
**Severity:** P1/P2/P3/P4
**Duration:** [start time] — [resolution time]
**Impact:** [# users affected, # meetings missed]

**Timeline:**
- HH:MM — Issue first reported
- HH:MM — Triage and severity assigned
- HH:MM — Workaround communicated
- HH:MM — Root cause identified
- HH:MM — Fix applied / service restored

**Root Cause:** [description]
**Resolution:** [what fixed it]
**Prevention:** [what to change to prevent recurrence]
**Action Items:**
- [ ] [who] [what] [by when]

Output

• Incident triaged and severity assigned
• Remediation applied or workaround enabled
• Stakeholders notified with status updates
• Post-incident review documented with prevention actions

Error Handling

Examples

For a missing synthetic meeting record, document severity=P2; integration=staging-calendar; window=15m; action=reconnect-scoped-integration; recording_probe=pass; transcript_retention=none; rollback=not-needed. The receipt uses no attendee name, title, recording, or transcript text.

Resources

• Granola Status Page
• Help Center
• Transcription Troubleshooting
• Granola Updates (Known Issues)

Next Steps

Proceed to granola-data-handling for data export, retention, and GDPR compliance.