Skip to content

IDSolvara (own platform) · 2026

Running 14 apps on one server: Solvara platform & ops

One 4 vCPU / 8 GB VPS runs every Solvara product. A custom ops CLI, encrypted env, observability and a Lead Finder keep it fast and safe.

Founder · infrastructure & tooling

CI deploy jobs, down from 8–10 min
~6 sCI deploy jobs, down from 8–10 min
command to roll back any app
1command to roll back any app
regencies crawled by Lead Finder
514regencies crawled by Lead Finder

Context

The situation

Solvara runs five live SaaS products plus a company site, an AI service and internal tools — around 14 apps — on a single VPS behind one Caddy reverse proxy.

The problem. Deploys burned GitHub Actions minutes waiting on server builds, .env files lived in plaintext on the server, and there was no single view of errors, uptime and load across apps.

Approach

Key decisions

  1. 01

    An audited ops CLI

    solvara-ops gives one command for status, logs, deploy and rollback. CI uses per-app restricted keys that can only deploy their own app, and every command is audit-logged.

  2. 02

    Encrypted environment files

    Every app’s .env is encrypted with SOPS + age and applied from source control, with drift checks before each deploy.

  3. 03

    Observability you own

    Self-hosted GlitchTip, Umami and Uptime Kuma feed an executive dashboard alongside server load.

  4. 04

    Lead Finder for sales

    A RabbitMQ-backed crawler covers 514 regencies, streams progress over SSE and sends picked leads straight into the sales CRM.

Screenshot

Solvara Tech company site showing its live products

Outcome

Results

  • CI deploy jobs went from 8–10 minutes of waiting to about 6 seconds; builds run on the server.
  • Errors, uptime, analytics and server load visible in one place.
  • Cloudflare Turnstile on logins across products after a security hardening pass.

Next case study

AI Assist for a telehealth funnel builder

Openloop Health