Servers
A server is the foundation for everything else — email, websites, applications and databases all run on one. It is a real dedicated VM on DigitalOcean or Microsoft Azure, in your account, with full root access. What Simplewala provides is everything above it.
Ordering a server
From the portal, go to Orders → New Order. The form asks three things.
Provider and region
Pick a location close to your visitors or your team. EU: Frankfurt or Helsinki. US: New York or San Francisco. The region affects latency and data residency — it does not affect the price, which is the same everywhere.
Size
Sizes are described in vCPUs (processing) and RAM (memory). As a starting point:
| Workload | Minimum size |
|---|---|
| Email only, up to 5 mailboxes | 1 vCPU, 2 GB RAM |
| Small WordPress site | 1 vCPU, 2 GB RAM |
| WordPress and email together | 2 vCPU, 4 GB RAM |
| Several apps, or email-heavy use | 4 vCPU, 8 GB RAM |
The order form also asks what the server is for — web, application, database, file, game, email or general — before it asks which machine. Machines are then ranked against that workload's binding resource and labelled Recommended, Workable or Undersized, with the reason. The answer also decides the block volume size and which software is installed during priming, so a file server does not end up with an unused MySQL.
When in doubt, start small. Moving up later is a supported operation.
Confirm
Review the monthly price and place the order. Provisioning takes roughly two to five minutes — longer than a raw VM takes to boot, because the machine is also primed: Docker, nginx, the firewall, SSL and monitoring are all configured before it is handed over.
Server statuses
| Status | Meaning | What to do |
|---|---|---|
| Active | Running normally. The expected state. | Nothing. |
| Provisioning | Being created and primed. | Wait two to five minutes. |
| Error | Provisioning failed, or the machine is unreachable. | Contact support — do not try to repair it by hand. |
| Suspended | Temporarily powered off, usually a billing issue. | Check Billing, then contact support. |
| Terminated | Permanently decommissioned. | Everything on it is gone. |
Agent status
Separately from the server itself, each machine runs a management agent. Connected means the platform can run commands on it. Disconnected means it cannot — your services are probably still running fine, but the dashboard cannot change anything until the agent is back.
“The server is active but my site is down”
The server being up says nothing about what runs on it. In order of likelihood:
- The app container crashed — see App not starting.
- DNS is not pointing at the server — see Domain not resolving.
- The SSL certificate expired — see Front Door → SSL.
Connecting over SSH
You do not need SSH. Email, apps, domains, databases, backups and logs are all managed from the portal. SSH is there for when you want to run something the dashboard does not cover, inspect a file directly, or debug something that is not visible from above.
The server's IPv4 address is on its detail page in the portal.
ssh root@your-server-ip On Windows, the built-in Windows Terminal takes the same command; PuTTY also works.
Adding your key
- Portal → Profile → SSH Keys → Add SSH Key.
- Paste your public key —
~/.ssh/id_ed25519.pubor~/.ssh/id_rsa.pub. - Save. The key is installed into the server's authorised keys and the password prompt stops.
Two things worth knowing before you change anything by hand
- The platform does not see manual changes. If you edit an nginx vhost or a container's configuration over SSH, the dashboard still believes its own version, and the next operation that rewrites that file will overwrite you. Change it in the portal where the portal manages it.
- You can lock yourself out. Editing the firewall or removing authorised keys over SSH is entirely possible. If it happens, support can get in through the cloud provider's console.
Databases on your server
Most people never touch these. Deploying WordPress creates its database and wires the app to it; nothing further is required. This section is for developers who want direct access.
Servers → your server → Databases lists the database containers on the machine and their status. MySQL and MariaDB suit WordPress and most PHP applications; PostgreSQL suits custom development. Valkey, Kafka, RabbitMQ and Memcached are available too — see Databases for the full reference.
The connection string is on each database's detail page:
mysql://dbuser:password@localhost:3306/dbname
postgresql://dbuser:password@localhost:5432/dbname Use localhost when the app is on the same server.
Reaching a database from your own machine
Database ports are not open to the internet by default. The safe way in is an SSH tunnel:
# forward the server's MySQL to port 3307 locally
ssh -L 3307:localhost:3306 root@your-server-ip
# then point your database client at localhost:3307 If you need a genuinely external port — for a service that cannot tunnel — the portal can allocate one that is proxied through nginx rather than exposed raw. See Networking.
Monitoring and logs
Every server reports per-container CPU, memory, network and disk metrics, plus per-domain request stats, without anything being installed. Servers → the server → Monitoring shows the current figures; Logs shows container and nginx output with the log lines parsed rather than dumped.
For retention and querying beyond that — dashboards, alert rules, historical series — an optional observability stack (VictoriaMetrics for metrics, VictoriaLogs for logs) can be installed onto the server from Observability. It runs on your own machine and its storage counts against your volumes, which is the tradeoff: no per-metric bill, and no data leaving your server.
Access is through the infrastructure domain at /_obs/ under the same certificate as
everything else, or through an on-demand external port. Both are covered in
Networking.