- Add HTTPS_ENABLED, TLS_CERT_FILE, TLS_KEY_FILE environment variables
- Server automatically starts in HTTPS mode when configured
- Falls back to HTTP with warning if certs missing
- WebSocket origins automatically adjust for HTTPS
- Fully backward compatible - defaults to HTTP
- Documented in README and CONFIGURATION.md
Addresses kenrmayfield's request for HTTPS support
- Grant enhanced Audit permissions to PBS tokens for system stats access
- Create PBSCard component to display CPU, memory, disk, uptime metrics
- Detect Docker PBS instances and show appropriate info without stats
- Add adaptive node layout on dashboard for better scalability
- Move PBS status card to Backups tab above frequency chart
- Create CompactNodeCard for handling many nodes efficiently
- Update PBS setup script to grant Audit role to both user and token
This addresses the request in issue #259 to show PBS system stats on the dashboard.
The implementation uses read-only Audit permissions to access the /nodes endpoint
while maintaining security. Docker PBS instances are detected and handled gracefully
since they cannot provide host system statistics.
- Only check cluster status during initial configuration, not during polling
- Cache cluster membership in config to avoid repeated API calls
- Skip cluster/resources endpoint entirely for standalone nodes
- Change cluster detection failure from WARN to DEBUG (expected for standalone)
This addresses #322 where standalone PVE nodes were causing certificate
lookup errors in syslog every minute during polling.
Requires 3 consecutive offline polls (~15 seconds) before triggering connectivity alerts to prevent false positives from transient cluster communication issues
addresses #329 - VMs were being displayed as LXC containers because the Type field wasn't being set when using the efficient cluster/resources polling method
Non-clustered Proxmox nodes were getting certificate verification errors
when Pulse tried to use the cluster/resources endpoint. Now checks if
the node is actually in a cluster before attempting efficient polling.
When both GetVersion and GetDatastores fail for PBS, properly set the
Status field to 'offline' and ConnectionHealth to 'error'. This prevents
the red dot from appearing when the instance state is undefined.
- Detect when running behind a proxy/tunnel (X-Forwarded-*, CF-Ray headers)
- Use SameSite=None for cookies when proxied to allow cross-origin access
- Properly detect HTTPS when behind proxy using X-Forwarded-Proto
- Fixes authentication not working through Cloudflare tunnels
The issue was that SameSite=Strict/Lax cookies don't work when the origin
changes (which happens with tunnels/proxies). Now we detect proxy headers
and relax the cookie policy to SameSite=None when needed.
- Mark deprecated poll functions that cause duplicate GetNodes() calls
- Add warnings when deprecated functions are called directly
- Previous fix only created WithNodes versions but didn't prevent the originals from being called
- This completes the fix started in commit fcd782370
- Reduces API calls and prevents certificate verification spam in syslog
The deprecated functions (pollVMs, pollContainers, pollStorage, pollStorageBackups)
still exist for backward compatibility but log warnings if called.
- Discord embeds don't support newlines in description field
- Use comma-separated list format for Discord grouped alerts
- Keep escaped newlines for other webhook providers (Telegram, Slack, Teams)
- Prevents JSON parsing errors with Discord webhook API
Discord now shows: "Alert | 🔔 5 alerts: • item1: 25.4%, • item2: 11.6%, ..."
Other providers show multi-line format with proper escaping.
- Show full list of grouped alerts for Discord, Slack, Teams, etc (not just Telegram)
- Properly escape newlines in grouped alert messages to prevent JSON parsing errors
- Ensure consistent formatting across all webhook providers
- Address issue where only custom templates showed complete alert lists
All webhook providers now display the same detailed grouped alert format with bullet points showing each alert's resource and value.
- Fixed double CPU percentage multiplication for containers/VMs
- Added CheckGuest calls to efficient polling path
- Fixed newline escaping in grouped webhook notifications
- Guest alerts now properly trigger for containers and VMs
These changes address issues where guest alerts weren't being triggered
at all due to the efficient polling path not calling CheckGuest, and
webhook notifications were failing due to unescaped newlines in grouped
alert messages breaking JSON templates.
Webhooks weren't being sent because the grouped alert message contained actual
newlines which broke JSON parsing in custom templates. Changed to use escaped
newlines (\n) which work properly in JSON strings.
- Fixed duplicate 'usage' text in storage alerts (Storage usage usage → Storage at X%)
- Improved grouped alerts to show details of other alerts instead of just count
- Added severity indicators (🔴 critical, 🟡 warning, 🟢 info)
- Better formatting with current value, threshold, type, duration, and timestamp
- Cleaner message structure with proper Markdown formatting
- Shows up to 3 additional alerts in grouped notifications with their values
- Fixed same template overwriting bug in sendGroupedWebhook function
- Grouped alerts now properly use custom templates
- Telegram webhooks now work for both individual and grouped alerts
- Successfully tested Telegram webhook delivery
- Login component now handles 429 rate limit responses correctly
- When rate limited, assume auth is configured and show login form
- /api/security/status endpoint excluded from strict auth rate limiting
- Status checks now use general API rate limit (500/min) instead of auth limit (10/min)
- Fixes issue where rapid logout/login could trigger rate limiting
- Fixes setup screen appearing incorrectly when rate limited
- Fixed bug where custom webhook templates were being overwritten with generic payload
- The issue occurred when a webhook had a custom template but empty service field
- Now properly preserves custom template payloads regardless of service field value
- Telegram webhooks and other custom template webhooks now work correctly
- Addresses user reports of webhook delivery issues
- Use realistic test values (85.5% CPU, 80% threshold) instead of zeros
- Set alert level to 'warning' instead of 'info' for better visibility
- Add 5-minute duration to test alerts for realistic testing
- Improve test message to be more descriptive
- Fix empty message text issue in Telegram webhooks
The webhook system now sends proper test notifications that accurately
simulate real alerts, making it easier to verify webhook configurations.
- Re-enable JSON validation that was accidentally commented out
- Fix Telegram chat_id handling with proper validation
- Add smart retry logic that distinguishes retryable from permanent errors
- Enhance error logging to include response bodies for debugging
- Add webhook URL validation for security (prevent SSRF)
- Fix deprecated strings.Title function usage
- Add webhook delivery history tracking
- Improve template variable substitution
- Add exponential backoff for retries (max 30 seconds)
- Validate numeric chat IDs for Telegram webhooks
- Add comprehensive error handling and logging
Tested with Telegram, Discord, and Slack - all working correctly now
- Fixed copy button not working in Quick Token Setup modal
- Button was conditionally rendered only when host was populated
- Now always visible and shows error if host is empty
- Updated placeholder text to be clearer
- Fixed toast notifications appearing behind modals
- Increased z-index from z-50 to z-[9999] to ensure visibility
- Fixed Generate API Token section intermittently not showing
- Added proper loading state for security status
- Prevents race condition where component renders before data loads
- Added loadSecurityStatus() function for reusability
- Refresh security status after password change
- Skip CSRF validation for /api/setup-script-url endpoint
- Endpoint generates temporary tokens, not a state change
- Fixes 403 Forbidden error when generating setup scripts
All API token functionality confirmed working with token rotation
- Cache nodes list in pollPVEInstance and pass to sub-functions
- Prevents multiple GetNodes() calls per polling cycle
- Reduces API calls from ~5 per cycle to 1 per cycle
- Fixes syslog spam on standalone PVE nodes trying to find cluster certificates
- Fixes PBS 'Transport endpoint not connected' errors from excessive polling
Previously we were calling GetNodes() in:
- pollPVEInstance (main)
- pollVMs
- pollContainers
- pollStorage
- pollStorageBackups
Now we call it once and pass the list to avoid duplicate API calls that trigger
certificate checks on non-clustered nodes.
- Fix alternating zero I/O metrics by implementing rate caching for stale data from Proxmox
- Hardcode polling interval to 10 seconds (matching Proxmox cluster/resources update cycle)
- Remove polling interval settings from UI (no longer user-configurable)
- Implement efficient VM/container polling using single cluster/resources API call
- Remove 'Remove Password' feature (auth is now mandatory)
- Fix CSRF validation for Basic Auth (exempt from CSRF checks)
- Fix Generate API Token modal and authentication
- Remove redundant 'Active' status from Authentication section
- Remove Connection Timeout setting from frontend (backend-only)
- Clean up frontend console logging (reduce verbosity)
- Remove PBS polling interval setting (fixed at 10s)
- Add frontend rebuild detection to backend-watch script
- Improve first-run setup flow and error handling
- Remove old /api/system/api-token endpoints
- Remove APIToken field from SystemSettings struct
- Remove token handling from system_settings.go
- Clean up config.Load() to not read token from system.json
- Remove unused frontend API token functions
- Remove unused APITokenManager and CurrentAPIToken components
API tokens are now managed exclusively via .env file with the new
/api/security/regenerate-token endpoint. This eliminates confusion
between the two systems and ensures consistency.
- Add backend endpoint to regenerate API tokens without resetting auth
- Updates .env file with new token while preserving other settings
- Frontend component with clear UX for token generation
- Shows new token once with copy functionality
- Indicates restart required to activate new token
- Works across all deployment types (Docker, LXC, native)
Much better UX than telling users to 'reconfigure security' just for a new token
- Add pending restart detection when .env exists but not loaded
- Update frontend to show pending state instead of re-showing setup
- Fix QuickSecuritySetup to refresh security status after configuration
- Remove auto-restart attempts from security setup
- Show deployment-appropriate restart instructions
- Update documentation to reflect new update mechanism
Related to security setup issues after removing sudo/auto-restart capabilities
Emergency release to fix critical issues in v4.3.7:
- Install script now correctly installs binary to /opt/pulse/bin/pulse
- Password changes no longer require sudo (addresses #317)
These fixes restore basic functionality for new installations
and Docker deployments.
- ProxmoxVE script can now check if auth is already set without credentials
- Quick-setup endpoint is now public but handler checks if setup should be skipped
- Prevents duplicate auth configuration attempts
- Critical for automated setup scripts
- Remove old buggy Quick Security Setup handler
- New handler in security_setup_fix.go handles all cases properly
- Tested and working with Docker, systemd, and manual installations
- Enhanced logging to diagnose chat_id extraction problems
- Log payload being sent to Telegram for debugging
- Better error messages when chat_id is missing or invalid
- Add clear error message for truncated bcrypt hashes (addresses #316)
- Detect hashes between 55-59 chars and warn they're invalid
- Only accept exactly 60 character bcrypt hashes as valid
- Treat truncated hashes as plaintext to force users to fix them
- Cluster now handles offline nodes gracefully without marking endpoints unhealthy
- Fixed error 595 (node unreachable) not being treated as node-specific failure
- Added parallel health checks with shorter timeouts for better performance
- Fixed inconsistent border width on offline node cards (removed conflicting border-l-4)
- Switched to ring utility for consistent outline on offline/alert nodes
- Improved logout functionality with proper CSRF token handling
addresses #312, #315
- alerts now trigger when nodes go offline or lose connectivity
- alerts automatically resolve when nodes come back online
- offline status is checked before resource thresholds
- logs critical events when nodes lose connectivity
This was a major oversight - nodes going offline should always trigger alerts
- adds logout button next to connection status indicator
- implements /api/logout endpoint to clear sessions
- button only shows when authentication is configured
- clears session cookie and invalidates server-side session
implements #315
- tracks online/offline status for individual cluster nodes
- updates ClusterEndpoint.Online field during node polling
- fixes issue where all cluster nodes showed green indicator regardless of status
fixes#312
addresses #314 - bcrypt hashes with $ signs were being interpreted as shell variables
by godotenv, causing authentication to fail after Docker container restarts
Script outputs from change-password and remove-password operations could
potentially contain sensitive information. Removed output logging while
keeping error logging for debugging.
Pulse Monitor