8.5 KiB

Raw Blame History

Authentication Session Timeout Fix - Deployment Guide

Overview

This document provides step-by-step instructions for deploying the authentication session timeout fixes that resolve the 2-minute logout issue.

Problem Summary

Users were experiencing unexpected logouts after exactly 2 minutes when navigating between pages. This was caused by:

Timing Race Condition: Authentication middleware cache expiry (2 minutes) and auth refresh plugin periodic validation (2 minutes) occurring simultaneously
No Request Deduplication: Multiple concurrent session checks causing conflicts
Insufficient Error Handling: Network errors triggering immediate logouts
No Grace Periods: Transient issues causing permanent session loss

Solution Overview

Core Changes

Session Manager Utility (server/utils/session-manager.ts)
- Centralized session management with request deduplication
- Promise caching for in-flight requests
- Network error grace periods
- Comprehensive logging and statistics
Authentication Middleware (middleware/authentication.ts)
- Changed cache expiry from 2 to 3 minutes with jitter
- Integrated SessionManager for deduplication
- Enhanced error handling and user feedback
Auth Refresh Plugin (plugins/01.auth-refresh.client.ts)
- Added random offset to prevent simultaneous validation
- Improved concurrent validation prevention
- Better error handling for network issues
Session API (server/api/auth/session.ts)
- Enhanced logging with request IDs
- Detailed error categorization
- Performance timing measurements
Keycloak Client (server/utils/keycloak-client.ts)
- Better error type distinction
- Increased retry attempts for token refresh
- Improved timeout handling
Refresh API (server/api/auth/refresh.ts)
- Enhanced error handling with request IDs
- Grace period support for transient failures
- Selective session clearing based on error type

Pre-deployment Checklist

Code Review: All changes reviewed and approved
Environment Variables: Verify all required environment variables are set
Dependencies: Confirm no new dependencies are required
Backup: Create backup of current production code
Monitoring: Ensure authentication logs are being captured
Testing: Verify fixes work in staging environment (if available)

Deployment Steps

Step 1: Deploy Session Manager Utility

Deploy server/utils/session-manager.ts
Verify no TypeScript compilation errors
Check server logs for any startup issues

Step 2: Update Authentication Middleware

Deploy updated middleware/authentication.ts
Monitor for any middleware errors in logs
Verify new timing configuration is active

Step 3: Update Auth Refresh Plugin

Deploy updated plugins/01.auth-refresh.client.ts
Check browser console for any client-side errors
Verify random offset is working (check logs)

Step 4: Update Session API

Deploy updated server/api/auth/session.ts
Monitor API endpoint logs for request IDs
Verify enhanced error messages are working

Step 5: Update Keycloak Client

Deploy updated server/utils/keycloak-client.ts
Check for any Keycloak communication errors
Verify retry logic is functioning

Step 6: Update Refresh API

Deploy updated server/api/auth/refresh.ts
Monitor token refresh operations
Verify graceful error handling

Post-deployment Verification

Immediate Verification (0-5 minutes)

No Deployment Errors

# Check server logs
tail -f /var/log/application.log | grep -E "(ERROR|FATAL)"

# Check for any 500 errors
curl -I https://your-domain.com/api/health

Login Flow Test
- Navigate to login page
- Complete authentication
- Verify successful redirect to dashboard

Session API Test

# Test session endpoint
curl -X GET https://your-domain.com/api/auth/session \
  -H "Cookie: nuxt-oidc-auth=<session-cookie>"

Short-term Verification (5-15 minutes)

Navigation Test
- Stay logged in for 5+ minutes
- Navigate between different pages
- Verify no unexpected logouts

Log Analysis

# Check for new session manager logs
grep "SESSION_MANAGER" /var/log/application.log

# Verify timing desynchronization
grep "Using cached session" /var/log/application.log

Long-term Verification (15+ minutes)

2-Minute Boundary Test
- Stay logged in for exactly 2 minutes
- Navigate to a new page
- Verify user remains authenticated
3-Minute Cache Test
- Stay on same page for 3+ minutes
- Navigate to new page
- Verify session is refreshed, not lost
Network Error Simulation
- Temporarily block network access
- Verify graceful degradation
- Restore network and verify recovery

Monitoring and Alerts

Key Metrics to Monitor

Authentication Errors

# Monitor auth failure rate
grep -c "AUTH_ERROR" /var/log/application.log

Session Manager Performance

# Check session check durations
grep "Session check completed" /var/log/application.log

Cache Hit Rate

# Monitor cache effectiveness
grep "Using cached session" /var/log/application.log | wc -l

Alert Thresholds

Auth Error Rate: > 5% of total auth checks
Session Check Duration: > 2 seconds average
Cache Miss Rate: > 80% (indicates caching issues)

Rollback Procedures

Immediate Rollback (if critical issues)

Stop Application
```
systemctl stop your-application
```

Restore Previous Code

git checkout previous-stable-tag
npm install
npm run build

Restart Application
```
systemctl start your-application
```
Verify Rollback
- Test login functionality
- Check error logs
- Verify user sessions work

Partial Rollback (if specific component issues)

Identify Problem Component
- Check which specific file is causing issues
- Review recent error logs

Rollback Specific Files

git checkout HEAD~1 -- middleware/authentication.ts
# or
git checkout HEAD~1 -- server/utils/session-manager.ts

Rebuild and Test

npm run build
systemctl restart your-application

Troubleshooting

Common Issues

Users Still Getting Logged Out at 2 Minutes
- Check if SessionManager is being used
- Verify cache expiry changes are active
- Look for timing synchronization issues
Session Check Errors
- Check network connectivity to Keycloak
- Verify environment variables are set
- Check Keycloak circuit breaker status
Performance Issues
- Monitor session check durations
- Check cache hit rates
- Verify request deduplication is working

Debug Commands

# Check session manager cache stats
curl https://your-domain.com/api/debug/session-cache-stats

# Monitor real-time auth logs
tail -f /var/log/application.log | grep -E "(SESSION|AUTH_REFRESH|MIDDLEWARE)"

# Check Keycloak connectivity
curl https://your-domain.com/api/debug/test-keycloak-connectivity

Success Criteria

The deployment is considered successful when:

No 2-Minute Logouts: Users can navigate freely after 2 minutes
Improved Error Handling: Network issues don't cause immediate logouts
Better Performance: Session checks complete faster due to caching
Enhanced Logging: Detailed logs help with debugging future issues
Graceful Degradation: System handles transient failures elegantly

Contact Information

For issues or questions regarding this deployment:

Technical Lead: [Your Name]
Emergency Contact: [Emergency Number]
Documentation: This file and related docs in /docs/ directory

Appendix

Environment Variables Required

KEYCLOAK_CLIENT_SECRET=your-secret-key
COOKIE_DOMAIN=.portnimara.dev

Log Examples

Successful session check:

[SESSION_MANAGER:abc123] Session check completed: {"authenticated":true,"reason":null,"fromCache":false}

Cache hit:

[SESSION_MANAGER:def456] Using cached session (age: 45 seconds)

Network error with grace period:

[SESSION_MANAGER:ghi789] Using cached result due to network error

Performance Benchmarks

Session Check Duration: < 500ms average
Cache Hit Rate: > 70%
Authentication Success Rate: > 99%
Network Error Recovery: < 5 seconds

8.5 KiB Raw Blame History