A curated guide to code sandboxing solutions, covering technologies like MicroVMs, application kernels, language runtimes, and containerization. It provides a feature matrix, in-depth platform profiles (e2b, Daytona, microsandbox, WebContainers, Replit, Cloudflare Workers, Fly.io, Kata Containers), and a decision framework for choosing the right sandboxing solution based on security, performance, workload type, and hosting preferences.
   
    
 
 
  
   
   Running GenAI models is easy. Scaling them to thousands of users, not so much. This guide details avenues for scaling AI workloads from proofs of concept to production-ready deployments, covering API integration, on-prem deployment considerations, hardware requirements, and tools like vLLM and Nvidia NIMs.
   
    
 
 
  
   
   K8S-native cluster-wide deployment for vLLM. Provides a reference implementation for building an inference stack on top of vLLM, enabling scaling, monitoring, request routing, and KV cache offloading with easy cloud deployment.
   
    
 
 
  
   
   vLLM Production Stack provides a reference implementation on how to build an inference stack on top of vLLM, allowing for scalable, monitored, and performant LLM deployments using Kubernetes and Helm.
   
    
 
 
  
   
   A user is seeking advice on deploying a new server with 4x H100 GPUs (320GB VRAM) for on-premise AI workloads. They are considering a Kubernetes-based deployment with RKE2, Nvidia GPU Operator, and tools like vLLM, llama.cpp, and Litellm. They are also exploring the option of GPU pass-through with a hypervisor. The post details their current infrastructure and asks for potential gotchas or best practices.
   
    
 
 
  
   
   Flagger is a progressive delivery Kubernetes operator that automates the release process for applications running on Kubernetes. It reduces the risk of introducing a new software version in production by gradually shifting traffic to the new version while measuring metrics and running conformance tests.
   
    
 
 
  
   
   This is a hands-on guide with Python example code that walks through the deployment of an ML-based search API using a simple 3-step approach. The article provides a deployment strategy applicable to most machine learning solutions, and the example code is available on GitHub.
   
    
 
 
  
   
   In this article, we explore how to deploy and manage machine learning models using Google Kubernetes Engine (GKE), Google AI Platform, and TensorFlow Serving. We will cover the steps to create a machine learning model and deploy it on a Kubernetes cluster for inference.
   
    
 
 
  
   
   apiVersion: v1
kind: Namespace
metadata:
  name: "mailserver"
---
apiVersion: "v1"
kind: "PersistentVolumeClaim"
metadata:
  name: "mailserver-pvc"
  namespace: "mailserver"
  labels:
    app: mailserver
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 15Gi
---
apiVersion: "v1"
kind: "PersistentVolumeClaim"
metadata:
  name: "mailserver-data-pvc"
  namespace: "mailserver"
  labels:
    app: mailserver
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 15Gi
---
kind: Service
apiVersion: v1
metadata:
  name: mailserver
  namespace: mailserver
  labels:
    app: mailserver
  annotations:
    metallb.universe.tf/address-pool: bgp-range
spec:
  selector:
    app: mailserver
  ports:
    - name: smtp
      port: 25
      targetPort: smtp
    - name: smtp-auth
      port: 587
      targetPort: smtp-auth
    - name: imap-secure
      port: 993
      targetPort: imap-secure
  type: LoadBalancer
  externalTrafficPolicy: Local
  loadBalancerIP: 192.168.4.2
---
kind: ConfigMap
apiVersion: v1
metadata:
  name: mailserver.config
  namespace: mailserver
  labels:
    app: mailserver
data:
  postfix-receive-access.cf: |
    blacklisted@recipient.example REJECT
  postfix-send-access.cf: |
    blacklisted-sender REJECT
  postfix-accounts.cf: |
    user1@example.com|{SHA512-CRYPT}$6$2YpW1nYtPBs2yLYS$z.5PGH1OEzsHHNhl3gJrc3D.YMZkvKw/vp.r5WIiwya6z7P/CQ9GDEJDr2G2V0cAfjDFeAQPUoopsuWPXLk3u1
  postfix-virtual.cf: |
    @example.org user1@example.com
  SigningTable: |
    *@example.org my._domainkey.example.org
    *@example.com my._domainkey.example.com
  KeyTable: |
    my._domainkey.example.org example.org:my:/etc/opendkim/keys/example.org-mail.key
    my._domainkey.example.com example.com:my:/etc/opendkim/keys/example.com-mail.key
  TrustedHosts: |
    127.0.0.1
    localhost
  spamassassin-rules.cf: |
    header SPAMMER_1  From =~ / a-z » +d{2,4}.com/i
    score SPAMMER_1  5.0
  sa-cron: |
    # This assumes you're having `environment: ONE_DIR=1` in the docker-compose.yml,
    # with a consolidated config in `/var/mail-state`
    #
    # m h dom mon dow user command
    #
    # Everyday 2:00AM, learn spam from a specific user
    # spam: junk directory
    0  2 * * * root  sa-learn --spam /var/mail/*/*/.Junk --dbpath /var/mail-state/lib-amavis/.spamassassin
    # ham: archive directories
    15 2 * * * root  sa-learn --ham /var/mail/*/*/.Archive* --dbpath /var/mail-state/lib-amavis/.spamassassin
    # ham: inbox subdirectories
    30 2 * * * root  sa-learn --ham /var/mail/*/*/cur* --dbpath /var/mail-state/lib-amavis/.spamassassin
---
kind: Secret
apiVersion: v1
metadata:
  name: mailserver.opendkim.keys
  namespace: mailserver
  labels:
    app: mailserver
type: Opaque
data:
  example.org-mail.key: "key here"
  example.com-mail.key: "key here"
---
kind: Deployment
apiVersion: apps/v1
metadata:
  name: mailserver
  namespace: mailserver
  labels:
    app: mailserver
spec:
  selector:
    matchLabels:
      app: mailserver
  strategy:
    type: Recreate
  template:
    metadata:
      labels:
        app: mailserver
    spec:
      nodeSelector:
        has/mail-server: "true"
      subdomain: mailserver
      containers:
        - name: mailserver
          image: radicand/docker-mailserver:gha
          # image: radicand/docker-mailserver:b9f43ab
          resources:
            requests:
              memory: "500Mi"
              cpu: "750m"
            limits:
              memory: "1500Mi"
              cpu: "900m"
          imagePullPolicy: Always
          securityContext: {}
          ports:
            - name: smtp
              containerPort: 25
            - name: smtp-auth
              containerPort: 587
            - name: imap-secure
              containerPort: 993
          env:
            - name: DOMAINNAME
              value: "example.com"
            - name: ONE_DIR
              value: "1"
            - name: SMTP_ONLY
              value: "0"
            - name: ENABLE_SRS
              value: "0"
            - name: ENABLE_SPAMASSASSIN
              value: "1"
            - name: SA_SPAM_SUBJECT
              value: " SPAM »  "
            - name: ENABLE_FAIL2BAN
              value: "0"
            - name: ENABLE_POSTGREY
              value: "0"
            - name: ENABLE_CLAMAV
              value: "0"
            - name: ENABLE_SASLAUTHD
              value: "0"
            - name: SRS_EXCLUDE_DOMAINS
              value: example.com
            - name: POSTMASTER_ADDRESS
              value: "postmaster@example.com"
            - name: POSTFIX_MESSAGE_SIZE_LIMIT
              value: "25000000"
            - name: SSL_TYPE
              value: "manual"
            - name: SSL_CERT_PATH
              value: "/etc/ssl/mailserver/tls.crt"
            - name: SSL_KEY_PATH
              value: "/etc/ssl/mailserver/tls.key"
            - name: OVERRIDE_HOSTNAME
              value: "mail.example.com"
            - name: REPORT_RECIPIENT
              value: "postmaster@example.com"
          volumeMounts:
            - name: config
              subPath: postfix-accounts.cf
              mountPath: /tmp/docker-mailserver/postfix-accounts.cf
              readOnly: true
            - name: config
              subPath: postfix-main.cf
              mountPath: /tmp/docker-mailserver/postfix-main.cf
              readOnly: true
            - name: config
              subPath: postfix-receive-access.cf
              mountPath: /tmp/docker-mailserver/postfix-receive-access.cf
              readOnly: true
            - name: config
              subPath: postfix-virtual.cf
              mountPath: /tmp/docker-mailserver/postfix-virtual.cf
              readOnly: true
            - name: config
              subPath: postfix-send-access.cf
              mountPath: /tmp/docker-mailserver/postfix-send-access.cf
              readOnly: true
            - name: config
              subPath: spamassassin-rules.cf
              mountPath: /tmp/docker-mailserver/spamassassin-rules.cf
              readOnly: true
            - name: config
              subPath: sa-cron
              mountPath: /etc/cron.d/sa-learn
              readOnly: true
              # mode: 0644
            - name: config
              subPath: SigningTable
              mountPath: /tmp/docker-mailserver/opendkim/SigningTable
              readOnly: true
            - name: config
              subPath: KeyTable
              mountPath: /tmp/docker-mailserver/opendkim/KeyTable
              readOnly: true
            - name: config
              subPath: TrustedHosts
              mountPath: /tmp/docker-mailserver/opendkim/TrustedHosts
              readOnly: true
            - name: opendkim-keys
              mountPath: /tmp/docker-mailserver/opendkim/keys
              readOnly: true
            - name: state
              mountPath: /var/mail-state
            - name: data
              mountPath: /var/mail
            - name: tls
              mountPath: /etc/ssl/mailserver
              readOnly: true
      volumes:
        - name: config
          configMap:
            name: mailserver.config
            # items:
            #   - key: sa-cron
            #     path: /etc/cron.d
            #     mode: 0644
        - name: opendkim-keys
          secret:
            secretName: mailserver.opendkim.keys
        - name: state
          persistentVolumeClaim:
            claimName: "mailserver-pvc"
        - name: data
          persistentVolumeClaim:
            claimName: "mailserver-data-pvc"
        - name: tls
          secret:
            secretName: mail-example-com-tls