StackBlaze vs a VPS
A droplet or an EC2 box gives you a kernel and an IP address. Everything that makes it a platform, deploys, TLS, databases, backups, monitoring, scaling, is a job you just took. StackBlaze ships that layer, from $5/mo.
StackBlaze
git push origin main
- Build, TLS, deploy on push
- Databases with backups
- Monitoring included
A VPS
apt upgrade && reboot
vim /etc/nginx/nginx.conf
certbot renew --force
- You are the platform team
- Patches never stop
- No rollbacks, no previews
The server is the easy part
The part under the water
On a VPS, you are the platform team
The server is the easy part. What you actually sign up for is six ongoing jobs, none of which ship your product.
Server setup
- ·Install and configure the OS
- ·Nginx or Caddy as reverse proxy
- ·SSL certificates and renewals
- ·SSH keys, users, sudo rules
Push to deploy. TLS, routing, and a public URL are automatic.
App environment
- ·Runtimes: Node, Python, Go, Ruby
- ·Process manager: PM2 or systemd
- ·Databases installed by hand
- ·Secrets in dotfiles on disk
Runtime auto-detected or Dockerfile. Databases are managed add-ons. Secrets are encrypted.
Security
- ·OS patches, forever
- ·Firewall rules, fail2ban
- ·DDoS and bot mitigation
- ·Dependency and image scanning
Platform patches the hosts. App Shield WAF and Trivy image scans are built in.
Monitoring
- ·Prometheus + Grafana to stand up
- ·Log aggregation to wire in
- ·Alerting rules and paging
- ·Uptime checks, somewhere else
CPU, memory, response time, throughput, and logs on every app, with AI root cause on incidents.
Scaling
- ·Resize the box, take the downtime
- ·Load balancer for a second node
- ·Session stickiness, connection pools
- ·Capacity planning by guesswork
Autoscaling between a min and max, replicas behind a load balancer, scale to zero when idle.
Deploys
- ·CI/CD pipeline you script yourself
- ·Rollback procedure you hope works
- ·Staging server no.2 to maintain
- ·Migrations run over SSH
Deploy on push, one-click rollbacks, preview environments per pull request on Pro.
Feature comparison
| Feature | StackBlaze | A VPS |
|---|---|---|
| Time to first deploy | Minutes | Hours to days |
| OS patches & hardening | Platform-managed | You, forever |
| SSL certificates | Automatic | certbot + renewals |
| Deploys & rollbacks | Built-in, one click | Scripts you write |
| Preview environments | Pro and above | |
| Managed Postgres + PITR | Yes (CloudNativePG) | Install & back up yourself |
| DB high availability | Operator failover | Manual replication |
| Monitoring & logs | Built-in | Prometheus/Grafana DIY |
| Autoscaling | Resize + downtime | |
| Scale to zero when idle | ||
| WAF & security scanning | App Shield + Trivy | Assemble yourself |
| Shell access | Console + SSH bastion | Root SSH |
| Root over the kernel | ||
| Monthly cost | From $5, spend cap | $5-$50 + your hours |
| Who gets paged at 2am | We do | You do |
Why StackBlaze
Databases run by Kubernetes operators, not containers on a volume
StackBlaze is built on Kubernetes, and every database add-on is managed by a battle-tested operator: CloudNativePG for Postgres, the MariaDB operator for MySQL, Strimzi for Kafka. You get the deployment, backup, and failover automation platform teams build in-house, without ever touching kubectl.
Scheduled backups, your retention
Postgres, MariaDB, and Valkey add-ons back up to object storage on a schedule you set, with a retention policy (7d, 4w, 3m) enforced by the operator.
Point-in-time recovery
Postgres archives WAL continuously; MariaDB takes physical backups. Restore to a timestamp as a new instance while the original keeps serving traffic.
HA with automatic failover
Scale a database to a replicated cluster and the operator handles failover. MariaDB promotion even rebinds ProxySQL so apps reconnect to the new primary.
Next to your app, privately
Every database runs on the same private network as your services, reachable over cluster DNS. No public endpoints, no cross-vendor egress fees.
On a VPS, the database is the scariest thing on the box: installed by hand, backed up by a cron script you wrote at midnight, restored never-tested. StackBlaze databases are operator-run clusters with scheduled backups, point-in-time restore, and failover included.
Postgres (CloudNativePG) · MySQL / MariaDB · Valkey (Redis) · MongoDB-compatible · Kafka (Strimzi) · RabbitMQ · ClickHouse · OpenSearch · CockroachDB · Cassandra · ScyllaDB · Milvus · Weaviate · S3-compatible storage
The other common question
“Why not just use AWS?”
AWS sells excellent parts. It does not sell an assembled platform. There are two ways to build on it, and both put you back in the operations business.
The EC2 path
An EC2 instance is a VPS with an IAM login. Every chore above still applies, plus security groups, AMI updates, and metered egress. You traded a $6 droplet for a pricier box and kept all the work.
The managed path
ECS or EKS, RDS, ElastiCache, ALB, ECR, CloudWatch, VPC, IAM, Secrets Manager. Real managed services, but you are now the integrator of nine of them, each with its own pricing model, and EKS bills for the control plane before you run a single container.
StackBlaze is that assembly already done: Kubernetes under the hood, with builds, TLS, databases, backups, monitoring, and preview environments wired together. One bill, a spend cap, and the same platform from the $5 Starter to a dedicated cluster.
What the VPS actually costs
The monthly server fee is the visible part. The rest of the bill arrives as tooling subscriptions and engineering time.
The sticker price
$6/mo
A droplet or small EC2 instance looks unbeatable. That price buys a kernel and an IP address. Everything else on this page is extra.
The tools you add anyway
+3-4 bills
Monitoring, log aggregation, backup storage, uptime checks. The tooling a platform includes arrives as separate subscriptions, each with its own invoice.
The expensive part
Your hours
Patch windows, certificate renewals, disk-full incidents, security advisories. At any realistic value for engineering time, the hours dwarf the server bill.
Like the VPS economics? Take a whole cluster, flat fee.
The honest appeal of a VPS is a fixed price for known hardware. StackBlaze Dedicated keeps that: a single-tenant HA Kubernetes cluster with reserved CPU and RAM for one flat monthly fee, with the platform still doing the patching, backups, and paging.
Dedicated
$300/mo
12 vCPU / 24 GB
3-node HA cluster, static egress IP
Dedicated Pro
$600/mo
24 vCPU / 48 GB
3-node HA cluster, 99.9% uptime SLA
Migrate off your VPS
Keep the server running while you move. Flip DNS when you're confident, cancel the box when you're done.
Connect the repository you already deploy from
Link GitHub (or Gitea, GitLab, Bitbucket) and pick the repo. StackBlaze auto-detects the runtime and build, or uses the Dockerfile you already have on the server.
Recreate your environment variables
Copy the .env file off the box and paste it into the StackBlaze variable editor. Same key/value format, stored encrypted instead of sitting on disk.
Provision databases and import your data
Create a managed Postgres, MySQL, or Redis add-on and restore a dump (pg_dump | pg_restore works as-is). Connection strings are injected automatically, and backups start on a schedule.
Move cron jobs and background workers
Crontab entries become scheduled jobs; long-running daemons become worker services. Both live next to your web app on the same private network.
Point DNS, then retire the box
Add your domain, update the CNAME, and let SSL provision automatically. Keep the VPS around as a fallback until traffic is verified, then cancel it and stop patching.
The questions everyone asks
Why not just use AWS?
AWS gives you excellent parts, not a platform. EC2 is a VPS with extra steps. Assembling the managed services (ECS, RDS, ALB, CloudWatch, IAM) is a real solution, and also a part-time job with nine pricing models. StackBlaze is that assembly already done, on Kubernetes, with one bill and a spend cap. When you want cloud-scale isolation, a Dedicated cluster gives you single-tenant hardware without the integration project.
Isn’t a $6 VPS cheaper?
The server is cheaper. The system is not. Add the monitoring, backup, and log tooling you would buy anyway, then the hours spent patching, hardening, and firefighting, and the droplet stops being cheap. StackBlaze Starter is $5/mo with a usage credit and a spend cap, and the ops work ships with it.
Do I lose SSH and control?
You keep a shell: every app has a web console and an SSH bastion (ssh.stackblaze.app), plus a file browser, live logs, and metrics. What you give up is root on a kernel, which in practice meant you were the one patching that kernel. Build with a Dockerfile if you want full control of the userland.
When is a VPS actually the right call?
When you need kernel modules, exotic networking, or hardware the platform does not offer, or when infrastructure is your product and you have the team to run it. For a middle path, StackBlaze Dedicated gives you a single-tenant HA cluster at a flat monthly fee: VPS-style cost certainty with the platform still doing the operations.
Am I locked in?
Less than on a hand-built server, honestly. Apps run as standard containers on standard Kubernetes; databases are stock engines (Postgres is Postgres, dump and restore anywhere); config exports as stackblaze.yaml and there is a Terraform provider and CLI. A three-year-old VPS with undocumented nginx configs is the harder thing to leave.
Keep the shell. Lose the pager.
Starter is $5/mo with a usage credit and a spend cap. Your first deploy takes minutes, and nobody has to patch it on Saturday.