Edit Nginx and Caddy Routes Without SSH

A reverse proxy usually becomes untouchable the moment it starts serving real traffic. You may have added one route months ago, tuned a header for WebSockets, or changed an upstream after moving a container. When the next adjustment arrives, you remember the risky part: SSH into the server, find the right file, edit it with production traffic flowing, run the provider’s validation command, and hope the reload does not disturb every other site on the box.

The route may have started in a visual workflow, but the first custom change sends you back to a terminal. Worse, a filename that looks familiar does not tell you who owns it. Editing the wrong Caddy fragment or Nginx vhost can turn a small routing change into an outage—or let an automation tool overwrite configuration you intended to manage yourself.

What changed

Edit Managed Proxy Routes lets you adjust a Server Compass-owned Nginx or Caddy route from the domain’s Config tab, without leaving the app. You edit the actual per-domain fragment, while Server Compass handles the backup, full configuration check, service reload, rollback, and live route verification around it.

How it works in practice

Start from the route you already manage

Open a server, choose Domain, and select the domain you want to change. An external Nginx or Caddy route now has four focused tabs: Overview, SSL, Config, and Advanced.

Overview gives you the context before you edit. You can see the current upstream, container port, stack, proxy provider, certificate state, route owner, and every managed file path attached to the domain. That means you can confirm you are looking at the right route before touching its configuration.

Move to Config and the Applied configuration panel lists the fragment Server Compass created. For Nginx, editable paths are restricted to Server Compass files under sites-available or conf.d. For Caddy, the editor accepts only per-domain files under /etc/caddy/server-compass/. Main provider files and vhosts you created yourself are outside the editor’s reach.

Edit the real fragment in a focused editor

Click Edit beside the managed path. The modal is labelled Edit Nginx Configuration or Edit Caddy Configuration, shows the filename, and loads the current content into a monospace editor. You are changing the configuration the provider actually reads, not a simplified form that may hide a directive you care about.

The editor tracks whether you have changed anything. Save & Reload Nginx or Save & Reload Caddy stays unavailable until the content differs, and closing with unsaved work asks you to confirm. That gives you room to inspect a route without accidentally writing it back.

This is useful when a normal reverse-proxy route needs one deliberate exception: a larger upload limit, a WebSocket header, a longer upstream timeout, a path-specific handler, or another directive your app requires. You keep the convenience of a managed route without losing access to the provider configuration beneath it.

Saving includes the checks you would normally run by hand

Clicking Save & Reload does more than replace a file. Server Compass uploads the edited content to a temporary path, copies the working fragment to a unique backup, and installs the new version with controlled permissions.

It then validates the complete proxy configuration—not only the fragment you changed. Nginx routes must pass nginx -t. Caddy routes must pass caddy validate --config /etc/caddy/Caddyfile. The active service is checked before any reload is attempted, so a stopped or unobservable proxy is not treated as a safe target.

Only after those checks pass does Server Compass reload Nginx or Caddy. If validation or reload fails, the previous fragment is restored, checked again, and reloaded. You see the error in the editor instead of discovering a broken route from a user report.

After a successful save, the domain is re-verified against the live server. Its status therefore reflects the route that is actually running, rather than merely confirming that a file was written.

Your existing configuration stays read-only

The editor appears only when the route owner is server_compass and the path matches the app’s strict per-domain naming rules. A panel-native route, an imported vhost, or a file you manage yourself remains read-only. Server Compass can report drift and show the provider, but it will not rewrite that route.

That ownership boundary matters most on servers you did not build from scratch. You can let Server Compass create and safely maintain new routes beside years of hand-managed Nginx or Caddy configuration, without granting it permission to edit everything the proxy serves.

When you add a domain through an existing proxy, the routing flow makes that choice explicit. You can keep Caddy in External Proxy mode, then create a route that Server Compass owns and can edit later; the rest of your Caddy setup remains yours.

Before vs After

Changing a managed proxy route	Before	Now
Find the right file	SSH in and search Nginx or Caddy directories	Open the domain’s Config tab and see the recorded managed path
Confirm ownership	Infer it from names and comments	Edit appears only for Server Compass-owned fragments
Make the change	Use a terminal editor over SSH	Edit the live fragment in a focused in-app editor
Protect the working route	Copy a backup manually	A unique backup is created before installation
Validate	Remember nginx -t or the Caddy validation command	The complete provider configuration is checked automatically
Reload	Run systemctl reload and inspect errors	The active service reloads only after validation passes
Recover a failure	Restore the old file and reload again	The previous fragment is restored and reloaded automatically
Confirm the result	Curl the domain and interpret the response	The live route is re-verified and its status refreshed

Who benefits most

Self-hosters with an established proxy. You can keep the Nginx or Caddy setup already serving your server, create new managed routes beside it, and make app-specific adjustments without turning every change into an SSH session.

Teams with applications that need custom proxy behavior. WebSockets, large uploads, streaming responses, and long-running requests often need one extra directive. You can make that precise change while keeping validation and recovery wrapped around the edit.

Anyone responsible for an inherited server. Ownership is visible before editing. Server Compass-owned fragments are editable; foreign configuration is read-only. You can improve one route without assuming control of every site on the machine.

Try it

Open a domain routed through external Nginx or Caddy, choose Config, and inspect the managed fragment. Click Edit, make the smallest change your app needs, and use Save & Reload. You get direct control over the route with the backup, validation, reload, rollback, and live verification steps handled in the same workflow.