Multiple Apps on One Server: Planning Guide | TinyPod

How many apps can one server handle? Plan CPU, memory, disk, and network for multi-app self-hosting deployments.

The Question

Can I run Nextcloud, Gitea, Grafana, Ghost, and Uptime Kuma on one server? Let's find out.

Memory Budget (The Usual Bottleneck)

Base System

OS + container runtime: ~500 MB

Databases (Shared)

PostgreSQL: 256 MB - 1 GB (depending on shared_buffers)

Redis: 50-100 MB

Applications (Typical)

Nextcloud: 256-512 MB

Gitea: 128-256 MB

Grafana: 128-256 MB

Ghost: 128-256 MB

Uptime Kuma: 128 MB

Plausible: 256-512 MB

Mattermost: 256-512 MB

Vaultwarden: 64-128 MB

Example: 4 GB Server

OS + runtime: 500 MB

PostgreSQL: 512 MB

Redis: 100 MB

Remaining: ~2.9 GB

Fits: 5-8 medium apps comfortably

CPU Budget

Most self-hosted apps are I/O bound, not CPU bound. They spend time waiting for database queries and network requests.

2 vCPU Server

Handles 5-10 apps easily unless they do compute-heavy work (image processing, AI inference, video transcoding).

Disk Budget

Container Images

Each app: 100 MB - 1 GB. Budget 5-10 GB for images.

Database Data

Varies wildly. 1-10 GB for most small deployments.

Application Data

User uploads, files, media. Budget based on use case.

Logs

1-5 GB if not configured with rotation.

Total

40 GB SSD comfortably holds a 5-8 app deployment.

Practical Setups

2 GB RAM, 1 vCPU, 20 GB

2-3 lightweight apps (blog, monitoring, password manager)

4 GB RAM, 2 vCPU, 40 GB

5-8 apps (the sweet spot for personal self-hosting)

8 GB RAM, 4 vCPU, 80 GB

10+ apps or resource-heavy apps (PostHog, GitLab, Keycloak)

Monitoring

Deploy first, then monitor with Grafana + Prometheus. If memory consistently exceeds 80%, upgrade the server. Don't over-provision upfront.

TinyPod's $5 server is the perfect starting point. Monitor, and scale up only when needed.

Running Multiple Apps on One Server: Resource Planning Guide