Updating Container Versions

This guide covers updating FoundationaLLM container images in Standard (AKS) deployments.

Overview

Container updates are required to:

Apply bug fixes
Access new features
Address security vulnerabilities
Maintain platform compatibility

Prerequisites

Requirement	Description
Azure CLI	Configured for correct tenant/subscription
AKS Credentials	In `.kube/config` directory
Helm CLI	Installed and configured
kubectl	Installed and configured

Getting AKS Credentials

az aks get-credentials `
    --name "<aks-cluster-name>" `
    --resource-group "<resource-group>"

Update Process

Step 1: Check Current Versions

# List all Helm releases
helm list -A

# Check specific deployment
kubectl get deployment -n fllm -o wide

Step 2: Get Current Values

Export the current Helm values:

# For backend services
helm get values "foundationallm" --all > gvalues.yaml

# For frontend services
helm get values "foundationallm-web" --all > gvalues-web.yaml

Step 3: Identify Target Version

Check available versions:

GitHub Releases
Container registry tags

Step 4: Update Images

Use the deployment script:

cd deploy/scripts

.\Deploy-Images-Aks.ps1 `
    -aksName "<aks-name>" `
    -resourceGroup "<resource-group>" `
    -tag "<version-tag>" `
    -charts "*" `
    -namespace "fllm"

Script Parameters

Parameter	Default	Description
`-name`	`foundationallm`	Helm release name
`-aksName`	(Required)	AKS cluster name
`-resourceGroup`	(Required)	Azure resource group
`-tag`	`latest`	Image version tag
`-charts`	`*`	Charts to deploy (`*` for all)
`-valuesFile`	`gvalues.yaml`	Helm values file path
`-namespace`	(Current)	Kubernetes namespace
`-tlsEnv`	`prod`	TLS environment (`prod`, `staging`, `none`, `custom`)
`-tlsHost`	-	Custom TLS hostname
`-tlsSecretName`	-	Custom TLS secret name
`-autoscale`	`$false`	Enable autoscaling

Selective Updates

Update Specific Charts

# Update only API services
.\Deploy-Images-Aks.ps1 `
    -aksName "<aks-name>" `
    -resourceGroup "<resource-group>" `
    -tag "0.9.0" `
    -charts "core-api,management-api"

Available Charts

Chart	Description
`core-api`	Core API service
`management-api`	Management API service
`orchestration-api`	Orchestration service
`gatekeeper-api`	Content safety service
`chat-ui`	Chat portal
`management-ui`	Management portal

Update Frontend Only

.\Deploy-Images-Aks.ps1 `
    -aksName "<frontend-aks-name>" `
    -resourceGroup "<resource-group>" `
    -tag "0.9.0" `
    -charts "chat-ui,management-ui"

Rolling Updates

Kubernetes performs rolling updates by default:

Old pods remain until new pods are healthy
No downtime during updates
Automatic rollback on failure

Monitor Update Progress

# Watch deployment status
kubectl rollout status deployment/core-api -n fllm

# View pod status
kubectl get pods -n fllm -w

Rollback if Needed

# Rollback to previous version
kubectl rollout undo deployment/core-api -n fllm

# Rollback Helm release
helm rollback foundationallm 1 -n fllm

Custom Values

Modifying Configuration

Edit gvalues.yaml to customize:

# Example: Increase replica count
coreApi:
  replicaCount: 3
  resources:
    requests:
      cpu: "500m"
      memory: "512Mi"
    limits:
      cpu: "1000m"
      memory: "1Gi"

Apply Custom Values

.\Deploy-Images-Aks.ps1 `
    -aksName "<aks-name>" `
    -resourceGroup "<resource-group>" `
    -tag "0.9.0" `
    -valuesFile "custom-values.yaml"

Verifying Updates

Check Image Versions

kubectl get pods -n fllm -o jsonpath="{.items[*].spec.containers[*].image}"

Check Service Health

# Get pod status
kubectl get pods -n fllm

# Check logs
kubectl logs -n fllm deployment/core-api --tail=100

Test Functionality

Access Chat Portal
Send a test message
Verify agent response
Check Management Portal

Troubleshooting

Issue	Solution
Pods not starting	Check `kubectl describe pod <pod-name> -n fllm`
Image pull errors	Verify registry access and image tag
Service unavailable	Check `kubectl get svc -n fllm`
Configuration issues	Compare `gvalues.yaml` with defaults

Common Errors

ImagePullBackOff:

# Check image details
kubectl describe pod <pod-name> -n fllm | grep -A 5 "Image:"

# Verify registry credentials
kubectl get secrets -n fllm

CrashLoopBackOff:

# Check pod logs
kubectl logs <pod-name> -n fllm --previous

# Check events
kubectl get events -n fllm --sort-by='.lastTimestamp'

Maintenance Windows

Best Practices

Schedule Updates during low-traffic periods
Test in Staging before production
Review Release Notes for breaking changes
Backup before major updates
Monitor closely after updates

Update Checklist

[ ] Review release notes
[ ] Backup current configuration
[ ] Update staging environment
[ ] Test functionality in staging
[ ] Schedule production maintenance window
[ ] Update production
[ ] Verify all services
[ ] Monitor for issues

Table of Contents