Adds docs.

docs/MAINTENANCE.md

@@ -0,0 +1,23 @@
+# Maintenance Guide
+
+Keep your wild cloud running smoothly.
+
+- [Security Best Practices](./guides/security.md)
+- [Monitoring](./guides/monitoring.md)
+- [Making backups](./guides/making-backups.md)
+- [Restoring backups](./guides/restoring-backups.md)
+
+## Upgrade
+
+- [Upgrade applications](./guides/upgrade-applications.md)
+- [Upgrade kubernetes](./guides/upgrade-kubernetes.md)
+- [Upgrade Talos](./guides/upgrade-talos.md)
+- [Upgrade Wild Cloud](./guides/upgrade-wild-cloud.md)
+
+## Troubleshooting
+
+- [Cluster issues](./guides/troubleshoot-cluster.md)
+- [DNS issues](./guides/troubleshoot-dns.md)
+- [Service connectivity issues](./guides/troubleshoot-service-connectivity.md)
+- [TLS certificate issues](./guides/troubleshoot-tls-certificates.md)
+- [Visibility issues](./guides/troubleshoot-visibility.md)

docs/guides/making-backups.md

@@ -0,0 +1,265 @@
+# Making Backups
+
+This guide covers how to create backups of your wild-cloud infrastructure using the integrated backup system.
+
+## Overview
+
+The wild-cloud backup system creates encrypted, deduplicated snapshots using restic. It backs up three main components:
+
+- **Applications**: Database dumps and persistent volume data
+- **Cluster**: Kubernetes resources and etcd state
+- **Configuration**: Wild-cloud repository and settings
+
+## Prerequisites
+
+Before making backups, ensure you have:
+
+1. **Environment configured**: Run `source env.sh` to load backup configuration
+2. **Restic repository**: Backup repository configured in `config.yaml`
+3. **Backup password**: Set in wild-cloud secrets
+4. **Staging directory**: Configured path for temporary backup files
+
+## Backup Components
+
+### Applications (`wild-app-backup`)
+
+Backs up individual applications including:
+- **Database dumps**: PostgreSQL/MySQL databases in compressed custom format
+- **PVC data**: Application files streamed directly for restic deduplication
+- **Auto-discovery**: Finds databases and PVCs based on app manifest.yaml
+
+### Cluster Resources (`wild-backup --cluster-only`)
+
+Backs up cluster-wide resources:
+- **Kubernetes resources**: All pods, services, deployments, secrets, configmaps
+- **Storage definitions**: PersistentVolumes, PVCs, StorageClasses  
+- **etcd snapshot**: Complete cluster state for disaster recovery
+
+### Configuration (`wild-backup --home-only`)
+
+Backs up wild-cloud configuration:
+- **Repository contents**: All app definitions, manifests, configurations
+- **Settings**: Wild-cloud configuration files and customizations
+
+## Making Backups
+
+### Full System Backup (Recommended)
+
+Create a complete backup of everything:
+
+```bash
+# Backup all components (apps + cluster + config)
+wild-backup
+```
+
+This is equivalent to:
+```bash
+wild-backup --home --apps --cluster
+```
+
+### Selective Backups
+
+#### Applications Only
+```bash
+# All applications
+wild-backup --apps-only
+
+# Single application  
+wild-app-backup discourse
+
+# Multiple applications
+wild-app-backup discourse gitea immich
+```
+
+#### Cluster Only
+```bash
+# Kubernetes resources + etcd
+wild-backup --cluster-only
+```
+
+#### Configuration Only
+```bash  
+# Wild-cloud repository
+wild-backup --home-only
+```
+
+### Excluding Components
+
+Skip specific components:
+
+```bash
+# Skip config, backup apps + cluster
+wild-backup --no-home
+
+# Skip applications, backup config + cluster  
+wild-backup --no-apps
+
+# Skip cluster resources, backup config + apps
+wild-backup --no-cluster
+```
+
+## Backup Process Details
+
+### Application Backup Process
+
+1. **Discovery**: Parses `manifest.yaml` to find database and PVC dependencies
+2. **Database backup**: Creates compressed custom-format dumps
+3. **PVC backup**: Streams files directly to staging for restic deduplication  
+4. **Staging**: Organizes files in clean directory structure
+5. **Upload**: Creates individual restic snapshots per application
+
+### Cluster Backup Process
+
+1. **Resource export**: Exports all Kubernetes resources to YAML
+2. **etcd snapshot**: Creates point-in-time etcd backup via talosctl
+3. **Upload**: Creates single restic snapshot for cluster state
+
+### Restic Snapshots
+
+Each backup creates tagged restic snapshots:
+
+```bash
+# View all snapshots
+restic snapshots
+
+# Filter by component
+restic snapshots --tag discourse    # Specific app
+restic snapshots --tag cluster      # Cluster resources
+restic snapshots --tag wc-home      # Wild-cloud config
+```
+
+## Where Backup Files Are Staged
+
+Before uploading to your restic repository, backup files are organized in a staging directory. This temporary area lets you see exactly what's being backed up and helps with deduplication.
+
+Here's what the staging area looks like:
+
+```
+backup-staging/
+├── apps/
+│   ├── discourse/
+│   │   ├── database_20250816T120000Z.dump
+│   │   ├── globals_20250816T120000Z.sql  
+│   │   └── discourse/
+│   │       └── data/         # All the actual files
+│   ├── gitea/
+│   │   ├── database_20250816T120000Z.dump
+│   │   └── gitea-data/
+│   │       └── data/         # Git repositories, etc.
+│   └── immich/
+│       ├── database_20250816T120000Z.dump
+│       └── immich-data/
+│           └── upload/       # Photos and videos
+└── cluster/
+    ├── all-resources.yaml    # All running services
+    ├── secrets.yaml          # Passwords and certificates
+    ├── configmaps.yaml       # Configuration data
+    └── etcd-snapshot.db      # Complete cluster state
+```
+
+This staging approach means you can examine backup contents before they're uploaded, and restic can efficiently deduplicate files that haven't changed.
+
+## Advanced Usage
+
+### Custom Backup Scripts
+
+Applications can provide custom backup logic:
+
+```bash
+# Create apps/myapp/backup.sh for custom behavior
+chmod +x apps/myapp/backup.sh
+
+# wild-app-backup will use custom script if present
+wild-app-backup myapp
+```
+
+### Monitoring Backup Status
+
+```bash
+# Check recent snapshots
+restic snapshots | head -20
+
+# Check specific app backups
+restic snapshots --tag discourse
+
+# Verify backup integrity
+restic check
+```
+
+### Backup Automation
+
+Set up automated backups with cron:
+
+```bash
+# Daily full backup at 2 AM
+0 2 * * * cd /data/repos/payne-cloud && source env.sh && wild-backup
+
+# Hourly app backups during business hours  
+0 9-17 * * * cd /data/repos/payne-cloud && source env.sh && wild-backup --apps-only
+```
+
+## Performance Considerations
+
+### Large PVCs (like Immich photos)
+
+The streaming backup approach provides:
+- **First backup**: Full transfer time (all files processed)
+- **Subsequent backups**: Only changed files processed (dramatically faster)
+- **Storage efficiency**: Restic deduplication reduces storage usage
+
+### Network Usage
+
+- **Database dumps**: Compressed at source, efficient transfer
+- **PVC data**: Uncompressed transfer, but restic handles deduplication
+- **etcd snapshots**: Small files, minimal impact
+
+## Troubleshooting
+
+### Common Issues
+
+**"No databases or PVCs found"**
+- App has no `manifest.yaml` with database dependencies
+- No PVCs with matching labels in app namespace
+- Create custom `backup.sh` script for special cases
+
+**"kubectl not found"** 
+- Ensure kubectl is installed and configured
+- Check cluster connectivity with `kubectl get nodes`
+
+**"Staging directory not set"**
+- Configure `cloud.backup.staging` in `config.yaml`
+- Ensure directory exists and is writable
+
+**"Could not create etcd backup"**
+- Ensure `talosctl` is installed for Talos clusters
+- Check control plane node connectivity
+- Verify etcd pods are accessible in kube-system namespace
+
+### Backup Verification
+
+Always verify backups periodically:
+
+```bash
+# Check restic repository integrity
+restic check
+
+# List recent snapshots
+restic snapshots --compact
+
+# Test restore to different directory
+restic restore latest --target /tmp/restore-test
+```
+
+## Security Notes
+
+- **Encryption**: All backups are encrypted with your backup password
+- **Secrets**: Kubernetes secrets are included in cluster backups
+- **Access control**: Secure your backup repository and passwords
+- **Network**: Consider bandwidth usage for large initial backups
+
+## Next Steps
+
+- [Restoring Backups](restoring-backups.md) - Learn how to restore from backups
+- Configure automated backup schedules
+- Set up backup monitoring and alerting
+- Test disaster recovery procedures

docs/guides/monitoring.md

@@ -0,0 +1,50 @@
+# System Health Monitoring
+
+## Basic Monitoring
+
+Check system health with:
+
+```bash
+# Node resource usage
+kubectl top nodes
+
+# Pod resource usage
+kubectl top pods -A
+
+# Persistent volume claims
+kubectl get pvc -A
+```
+
+## Advanced Monitoring (Future Implementation)
+
+Consider implementing:
+
+1. **Prometheus + Grafana** for comprehensive monitoring:
+   ```bash
+   # Placeholder for future implementation
+   helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
+   helm install prometheus prometheus-community/kube-prometheus-stack --namespace monitoring --create-namespace
+   ```
+
+2. **Loki** for log aggregation:
+   ```bash
+   # Placeholder for future implementation
+   helm repo add grafana https://grafana.github.io/helm-charts
+   helm install loki grafana/loki-stack --namespace logging --create-namespace
+   ```
+
+## Additional Resources
+
+This document will be expanded in the future with:
+
+- Detailed backup and restore procedures
+- Monitoring setup instructions
+- Comprehensive security hardening guide
+- Automated maintenance scripts
+
+For now, refer to the following external resources:
+
+- [K3s Documentation](https://docs.k3s.io/)
+- [Kubernetes Troubleshooting Guide](https://kubernetes.io/docs/tasks/debug/)
+- [Velero Backup Documentation](https://velero.io/docs/latest/)
+- [Kubernetes Security Best Practices](https://kubernetes.io/docs/concepts/security/)

docs/guides/restoring-backups.md

@@ -0,0 +1,294 @@
+# Restoring Backups
+
+This guide will walk you through restoring your applications and cluster from wild-cloud backups. Hopefully you'll never need this, but when you do, it's critical that the process works smoothly.
+
+## Understanding Restore Types
+
+Your wild-cloud backup system can restore different types of data depending on what you need to recover:
+
+**Application restores** bring back individual applications by restoring their database contents and file storage. This is what you'll use most often - maybe you accidentally deleted something in Discourse, or Gitea got corrupted, or you want to roll back Immich to before a bad update.
+
+**Cluster restores** are for disaster recovery scenarios where you need to rebuild your entire Kubernetes cluster from scratch. This includes restoring all the cluster's configuration and even its internal state.
+
+**Configuration restores** bring back your wild-cloud repository and settings, which contain all the "recipes" for how your infrastructure should be set up.
+
+## Before You Start Restoring
+
+Make sure you have everything needed to perform restores. You need to be in your wild-cloud directory with the environment loaded (`source env.sh`). Your backup repository and password should be configured and working - you can test this by running `restic snapshots` to see your available backups.
+
+Most importantly, make sure you have kubectl access to your cluster, since restores involve creating temporary pods and manipulating storage.
+
+## Restoring Applications
+
+### Basic Application Restore
+
+The most common restore scenario is bringing back a single application. To restore the latest backup of an app:
+
+```bash
+wild-app-restore discourse
+```
+
+This restores both the database and all file storage for the discourse app. The restore system automatically figures out what the app needs based on its manifest file and what was backed up.
+
+If you want to restore from a specific backup instead of the latest:
+
+```bash
+wild-app-restore discourse abc123
+```
+
+Where `abc123` is the snapshot ID from `restic snapshots --tag discourse`.
+
+### Partial Restores
+
+Sometimes you only need to restore part of an application. Maybe the database is fine but the files got corrupted, or vice versa.
+
+To restore only the database:
+```bash
+wild-app-restore discourse --db-only
+```
+
+To restore only the file storage:
+```bash
+wild-app-restore discourse --pvc-only
+```
+
+To restore without database roles and permissions (if they're causing conflicts):
+```bash
+wild-app-restore discourse --skip-globals
+```
+
+### Finding Available Backups
+
+To see what backups are available for an app:
+```bash
+wild-app-restore discourse --list
+```
+
+This shows recent snapshots with their IDs, timestamps, and what was included.
+
+## How Application Restores Work
+
+Understanding what happens during a restore can help when things don't go as expected.
+
+### Database Restoration
+
+When restoring a database, the system first downloads the backup files from your restic repository. It then prepares the database by creating any needed roles, disconnecting existing users, and dropping/recreating the database to ensure a clean restore.
+
+For PostgreSQL databases, it uses `pg_restore` with parallel processing to speed up large database imports. For MySQL, it uses standard mysql import commands. The system also handles database ownership and permissions automatically.
+
+### File Storage Restoration
+
+File storage (PVC) restoration is more complex because it involves safely replacing files that might be actively used by running applications.
+
+First, the system creates a safety snapshot using Longhorn. This means if something goes wrong during the restore, you can get back to where you started. Then it scales your application down to zero replicas so no pods are using the storage.
+
+Next, it creates a temporary utility pod with the PVC mounted and copies all the backup files into place, preserving file permissions and structure. Once the data is restored and verified, it removes the utility pod and scales your application back up.
+
+If everything worked correctly, the safety snapshot is automatically deleted. If something went wrong, the safety snapshot is preserved so you can recover manually.
+
+## Cluster Disaster Recovery
+
+Cluster restoration is much less common but critical when you need to rebuild your entire infrastructure.
+
+### Restoring Kubernetes Resources
+
+To restore all cluster resources from a backup:
+
+```bash
+# Download cluster backup
+restic restore --tag cluster latest --target ./restore/
+
+# Apply all resources
+kubectl apply -f restore/cluster/all-resources.yaml
+```
+
+You can also restore specific types of resources:
+```bash
+kubectl apply -f restore/cluster/secrets.yaml
+kubectl apply -f restore/cluster/configmaps.yaml
+```
+
+### Restoring etcd State
+
+**Warning: This is extremely dangerous and will affect your entire cluster.**
+
+etcd restoration should only be done when rebuilding a cluster from scratch. For Talos clusters:
+
+```bash
+talosctl --nodes <control-plane-ip> etcd restore --from ./restore/cluster/etcd-snapshot.db
+```
+
+This command stops etcd, replaces its data with the backup, and restarts the cluster. Expect significant downtime while the cluster rebuilds itself.
+
+## Common Disaster Recovery Scenarios
+
+### Complete Application Loss
+
+When an entire application is gone (namespace deleted, pods corrupted, etc.):
+
+```bash
+# Make sure the namespace exists
+kubectl create namespace discourse --dry-run=client -o yaml | kubectl apply -f -
+
+# Apply the application manifests if needed
+kubectl apply -f apps/discourse/
+
+# Restore the application data
+wild-app-restore discourse
+```
+
+### Complete Cluster Rebuild
+
+When rebuilding a cluster from scratch:
+
+First, build your new cluster infrastructure and install wild-cloud components. Then configure backup access so you can reach your backup repository.
+
+Restore cluster state:
+```bash
+restic restore --tag cluster latest --target ./restore/
+# Apply etcd snapshot using appropriate method for your cluster type
+```
+
+Finally, restore all applications:
+```bash
+# See what applications are backed up
+wild-app-restore --list
+
+# Restore each application individually
+wild-app-restore discourse
+wild-app-restore gitea
+wild-app-restore immich
+```
+
+### Rolling Back After Bad Changes
+
+Sometimes you need to undo recent changes to an application:
+
+```bash
+# See available snapshots
+wild-app-restore discourse --list
+
+# Restore from before the problematic changes
+wild-app-restore discourse abc123
+```
+
+## Cross-Cluster Migration
+
+You can use backups to move applications between clusters:
+
+On the source cluster, create a fresh backup:
+```bash
+wild-app-backup discourse
+```
+
+On the target cluster, deploy the application manifests:
+```bash
+kubectl apply -f apps/discourse/
+```
+
+Then restore the data:
+```bash
+wild-app-restore discourse
+```
+
+## Verifying Successful Restores
+
+After any restore, verify that everything is working correctly.
+
+For databases, check that you can connect and see expected data:
+```bash
+kubectl exec -n postgres deploy/postgres-deployment -- \
+  psql -U postgres -d discourse -c "SELECT count(*) FROM posts;"
+```
+
+For file storage, check that files exist and applications can start:
+```bash
+kubectl get pods -n discourse
+kubectl logs -n discourse deployment/discourse
+```
+
+For web applications, test that you can access them:
+```bash
+curl -f https://discourse.example.com/latest.json
+```
+
+## When Things Go Wrong
+
+### No Snapshots Found
+
+If the restore system can't find backups for an application, check that snapshots exist:
+```bash
+restic snapshots --tag discourse
+```
+
+Make sure you're using the correct app name and that backups were actually created successfully.
+
+### Database Restore Failures
+
+Database restores can fail if the target database isn't accessible or if there are permission issues. Check that your postgres or mysql pods are running and that you can connect to them manually.
+
+Review the restore error messages carefully - they usually indicate whether the problem is with the backup file, database connectivity, or permissions.
+
+### PVC Restore Failures
+
+If PVC restoration fails, check that you have sufficient disk space and that the PVC isn't being used by other pods. The error messages will usually indicate what went wrong.
+
+Most importantly, remember that safety snapshots are preserved when PVC restores fail. You can see them with:
+```bash
+kubectl get snapshot.longhorn.io -n longhorn-system -l app=wild-app-restore
+```
+
+These snapshots let you recover to the pre-restore state if needed.
+
+### Application Won't Start After Restore
+
+If pods fail to start after restoration, check file permissions and ownership. Sometimes the restoration process doesn't perfectly preserve the exact permissions that the application expects.
+
+You can also try scaling the application to zero and back to one, which sometimes resolves transient issues:
+```bash
+kubectl scale deployment/discourse -n discourse --replicas=0
+kubectl scale deployment/discourse -n discourse --replicas=1
+```
+
+## Manual Recovery
+
+When automated restore fails, you can always fall back to manual extraction and restoration:
+
+```bash
+# Extract backup files to local directory
+restic restore --tag discourse latest --target ./manual-restore/
+
+# Manually copy database dump to postgres pod
+kubectl cp ./manual-restore/discourse/database_*.dump \
+  postgres/postgres-deployment-xxx:/tmp/
+
+# Manually restore database
+kubectl exec -n postgres deploy/postgres-deployment -- \
+  pg_restore -U postgres -d discourse /tmp/database_*.dump
+```
+
+For file restoration, you'd need to create a utility pod and manually copy files into the PVC.
+
+## Best Practices
+
+Test your restore procedures regularly in a non-production environment. It's much better to discover issues with your backup system during a planned test than during an actual emergency.
+
+Always communicate with users before performing restores, especially if they involve downtime. Document any manual steps you had to take so you can improve the automated process.
+
+After any significant restore, monitor your applications more closely than usual for a few days. Sometimes problems don't surface immediately.
+
+## Security and Access Control
+
+Restore operations are powerful and can be destructive. Make sure only trusted administrators can perform restores, and consider requiring approval or coordination before major restoration operations.
+
+Be aware that cluster restores include all secrets, so they potentially expose passwords, API keys, and certificates. Ensure your backup repository is properly secured.
+
+Remember that Longhorn safety snapshots are preserved when things go wrong. These snapshots may contain sensitive data, so clean them up appropriately once you've resolved any issues.
+
+## What's Next
+
+The best way to get comfortable with restore operations is to practice them in a safe environment. Set up a test cluster and practice restoring applications and data.
+
+Consider creating runbooks for your most likely disaster scenarios, including the specific commands and verification steps for your infrastructure.
+
+Read the [Making Backups](making-backups.md) guide to ensure you're creating the backups you'll need for successful recovery.

docs/guides/troubleshoot-cluster.md

@@ -0,0 +1,19 @@
+# Troubleshoot Wild Cloud Cluster issues
+
+## General Troubleshooting Steps
+
+1. **Check Node Status**:
+   ```bash
+   kubectl get nodes
+   kubectl describe node <node-name>
+   ```
+
+1. **Check Component Status**:
+   ```bash
+   # Check all pods across all namespaces
+   kubectl get pods -A
+   
+   # Look for pods that aren't Running or Ready
+   kubectl get pods -A | grep -v "Running\|Completed"
+   ```
+

docs/guides/troubleshoot-dns.md

@@ -0,0 +1,20 @@
+# Troubleshoot DNS
+
+If DNS resolution isn't working properly:
+
+1. Check CoreDNS status:
+   ```bash
+   kubectl get pods -n kube-system -l k8s-app=kube-dns
+   kubectl logs -l k8s-app=kube-dns -n kube-system
+   ```
+
+2. Verify CoreDNS configuration:
+   ```bash
+   kubectl get configmap -n kube-system coredns -o yaml
+   ```
+
+3. Test DNS resolution from inside the cluster:
+   ```bash
+   kubectl run -i --tty --rm debug --image=busybox --restart=Never -- nslookup kubernetes.default
+   ```
+

docs/guides/troubleshoot-service-connectivity.md

@@ -0,0 +1,18 @@
+# Troubleshoot Service Connectivity
+
+If services can't communicate:
+
+1. Check network policies:
+   ```bash
+   kubectl get networkpolicies -A
+   ```
+
+2. Verify service endpoints:
+   ```bash
+   kubectl get endpoints -n <namespace>
+   ```
+
+3. Test connectivity from within the cluster:
+   ```bash
+   kubectl run -i --tty --rm debug --image=busybox --restart=Never -- wget -O- <service-name>.<namespace>
+   ```

docs/guides/troubleshoot-tls-certificates.md

@@ -0,0 +1,24 @@
+# Troubleshoot TLS Certificates
+
+If services show invalid certificates:
+
+1. Check certificate status:
+   ```bash
+   kubectl get certificates -A
+   ```
+
+2. Examine certificate details:
+   ```bash
+   kubectl describe certificate <cert-name> -n <namespace>
+   ```
+
+3. Check for cert-manager issues:
+   ```bash
+   kubectl get pods -n cert-manager
+   kubectl logs -l app=cert-manager -n cert-manager
+   ```
+
+4. Verify the Cloudflare API token is correctly set up:
+   ```bash
+   kubectl get secret cloudflare-api-token -n internal
+   ```

docs/guides/troubleshoot-visibility.md

@@ -0,0 +1,246 @@
+# Troubleshoot Service Visibility
+
+This guide covers common issues with accessing services from outside the cluster and how to diagnose and fix them.
+
+## Common Issues
+
+External access to your services might fail for several reasons:
+
+1. **DNS Resolution Issues** - Domain names not resolving to the correct IP address
+2. **Network Connectivity Issues** - Traffic can't reach the cluster's external IP
+3. **TLS Certificate Issues** - Invalid or missing certificates
+4. **Ingress/Service Configuration Issues** - Incorrectly configured routing
+
+## Diagnostic Steps
+
+### 1. Check DNS Resolution
+
+**Symptoms:**
+
+- Browser shows "site cannot be reached" or "server IP address could not be found"
+- `ping` or `nslookup` commands fail for your domain
+- Your service DNS records don't appear in CloudFlare or your DNS provider
+
+**Checks:**
+
+```bash
+# Check if your domain resolves (from outside the cluster)
+nslookup yourservice.yourdomain.com
+
+# Check if ExternalDNS is running
+kubectl get pods -n externaldns
+
+# Check ExternalDNS logs for errors
+kubectl logs -n externaldns -l app=external-dns  < /dev/null |  grep -i error
+kubectl logs -n externaldns -l app=external-dns | grep -i "your-service-name"
+
+# Check if CloudFlare API token is configured correctly
+kubectl get secret cloudflare-api-token -n externaldns
+```
+
+**Common Issues:**
+
+a) **ExternalDNS Not Running**: The ExternalDNS pod is not running or has errors.
+
+b) **Cloudflare API Token Issues**: The API token is invalid, expired, or doesn't have the right permissions.
+
+c) **Domain Filter Mismatch**: ExternalDNS is configured with a `--domain-filter` that doesn't match your domain.
+
+d) **Annotations Missing**: Service or Ingress is missing the required ExternalDNS annotations.
+
+**Solutions:**
+
+```bash
+# 1. Recreate CloudFlare API token secret
+kubectl create secret generic cloudflare-api-token \
+  --namespace externaldns \
+  --from-literal=api-token="your-api-token" \
+  --dry-run=client -o yaml | kubectl apply -f -
+
+# 2. Check and set proper annotations on your Ingress:
+kubectl annotate ingress your-ingress -n your-namespace \
+  external-dns.alpha.kubernetes.io/hostname=your-service.your-domain.com
+
+# 3. Restart ExternalDNS
+kubectl rollout restart deployment -n externaldns external-dns
+```
+
+### 2. Check Network Connectivity
+
+**Symptoms:**
+
+- DNS resolves to the correct IP but the service is still unreachable
+- Only some services are unreachable while others work
+- Network timeout errors
+
+**Checks:**
+
+```bash
+# Check if MetalLB is running
+kubectl get pods -n metallb-system
+
+# Check MetalLB IP address pool
+kubectl get ipaddresspools.metallb.io -n metallb-system
+
+# Verify the service has an external IP
+kubectl get svc -n your-namespace your-service
+```
+
+**Common Issues:**
+
+a) **MetalLB Configuration**: The IP pool doesn't match your network or is exhausted.
+
+b) **Firewall Issues**: Firewall is blocking traffic to your cluster's external IP.
+
+c) **Router Configuration**: NAT or port forwarding issues if using a router.
+
+**Solutions:**
+
+```bash
+# 1. Check and update MetalLB configuration
+kubectl apply -f infrastructure_setup/metallb/metallb-pool.yaml
+
+# 2. Check service external IP assignment
+kubectl describe svc -n your-namespace your-service
+```
+
+### 3. Check TLS Certificates
+
+**Symptoms:**
+
+- Browser shows certificate errors
+- "Your connection is not private" warnings
+- Cert-manager logs show errors
+
+**Checks:**
+
+```bash
+# Check certificate status
+kubectl get certificates -A
+
+# Check cert-manager logs
+kubectl logs -n cert-manager -l app=cert-manager
+
+# Check if your ingress is using the correct certificate
+kubectl get ingress -n your-namespace your-ingress -o yaml
+```
+
+**Common Issues:**
+
+a) **Certificate Issuance Failures**: DNS validation or HTTP validation failing.
+
+b) **Wrong Secret Referenced**: Ingress is referencing a non-existent certificate secret.
+
+c) **Expired Certificate**: Certificate has expired and wasn't renewed.
+
+**Solutions:**
+
+```bash
+# 1. Check and recreate certificates
+kubectl apply -f infrastructure_setup/cert-manager/wildcard-certificate.yaml
+
+# 2. Update ingress to use correct secret
+kubectl patch ingress your-ingress -n your-namespace --type=json \
+  -p='[{"op": "replace", "path": "/spec/tls/0/secretName", "value": "correct-secret-name"}]'
+```
+
+### 4. Check Ingress Configuration
+
+**Symptoms:**
+
+- HTTP 404, 503, or other error codes
+- Service accessible from inside cluster but not outside
+- Traffic routed to wrong service
+
+**Checks:**
+
+```bash
+# Check ingress status
+kubectl get ingress -n your-namespace
+
+# Check Traefik logs
+kubectl logs -n kube-system -l app.kubernetes.io/name=traefik
+
+# Check ingress configuration
+kubectl describe ingress -n your-namespace your-ingress
+```
+
+**Common Issues:**
+
+a) **Incorrect Service Targeting**: Ingress is pointing to wrong service or port.
+
+b) **Traefik Configuration**: IngressClass or middleware issues.
+
+c) **Path Configuration**: Incorrect path prefixes or regex.
+
+**Solutions:**
+
+```bash
+# 1. Verify ingress configuration
+kubectl edit ingress -n your-namespace your-ingress
+
+# 2. Check that the referenced service exists
+kubectl get svc -n your-namespace
+
+# 3. Restart Traefik if needed
+kubectl rollout restart deployment -n kube-system traefik
+```
+
+## Advanced Diagnostics
+
+For more complex issues, you can use port-forwarding to test services directly:
+
+```bash
+# Port-forward the service directly
+kubectl port-forward -n your-namespace svc/your-service 8080:80
+
+# Then test locally
+curl http://localhost:8080
+```
+
+You can also deploy a debug pod to test connectivity from inside the cluster:
+
+```bash
+# Start a debug pod
+kubectl run -i --tty --rm debug --image=busybox --restart=Never -- sh
+
+# Inside the pod, test DNS and connectivity
+nslookup your-service.your-namespace.svc.cluster.local
+wget -O- http://your-service.your-namespace.svc.cluster.local
+```
+
+## ExternalDNS Specifics
+
+ExternalDNS can be particularly troublesome. Here are specific debugging steps:
+
+1. **Check Log Level**: Set `--log-level=debug` for more detailed logs
+2. **Check Domain Filter**: Ensure `--domain-filter` includes your domain
+3. **Check Provider**: Ensure `--provider=cloudflare` (or your DNS provider)
+4. **Verify API Permissions**: CloudFlare token needs Zone.Zone and Zone.DNS permissions
+5. **Check TXT Records**: ExternalDNS uses TXT records for ownership tracking
+
+```bash
+# Restart with verbose logging
+kubectl set env deployment/external-dns -n externaldns -- --log-level=debug
+
+# Check for specific domain errors
+kubectl logs -n externaldns -l app=external-dns | grep -i yourservice.yourdomain.com
+```
+
+## CloudFlare Specific Issues
+
+When using CloudFlare, additional issues may arise:
+
+1. **API Rate Limiting**: CloudFlare may rate limit frequent API calls
+2. **DNS Propagation**: Changes may take time to propagate through CloudFlare's CDN
+3. **Proxied Records**: The `external-dns.alpha.kubernetes.io/cloudflare-proxied` annotation controls whether CloudFlare proxies traffic
+4. **Access Restrictions**: CloudFlare Access or Page Rules may restrict access
+5. **API Token Permissions**: The token must have Zone:Zone:Read and Zone:DNS:Edit permissions
+6. **Zone Detection**: If using subdomains, ensure the parent domain is included in the domain filter
+
+Check CloudFlare dashboard for:
+
+- DNS record existence
+- API access logs
+- DNS settings including proxy status
+- Any error messages or rate limit warnings

docs/guides/upgrade-applications.md

@@ -0,0 +1,3 @@
+# Upgrade Applications
+
+TBD

docs/guides/upgrade-kubernetes.md

@@ -0,0 +1,3 @@
+# Upgrade Kubernetes
+
+TBD

docs/guides/upgrade-talos.md

@@ -0,0 +1,3 @@
+# Upgrade Talos
+
+TBD

docs/guides/upgrade-wild-cloud.md

@@ -0,0 +1,3 @@
+# Upgrade Wild Cloud
+
+TBD