Server Requirements
This page describes hardware specifications for hosting CoderFlow. Three tiers are provided as starting points; pick the one that matches your expected workload. All tiers assume a Linux distribution (Ubuntu Server LTS is recommended) with Docker as the container runtime.
Configuration Tiers
| Component | Light Workload | Medium Workload | Heavy Workload |
|---|---|---|---|
| CPU | 4–8 vCPUs | 8–12 vCPUs | 16+ vCPUs |
| RAM | 16 GB | 32 GB | 64 GB+ |
| Storage Type | SSD (SATA or NVMe) | NVMe SSD | NVMe SSD (RAID optional) |
| Storage Size | 256–512 GB | 512 GB – 1 TB | 1 TB+ |
| Network | 1 Gbps | 1–10 Gbps | 10 Gbps+ |
| OS | Linux | Linux | Linux |
Critical Requirements
Storage
SSD storage is required. Spinning disks (HDDs) create severe I/O bottlenecks for container image pulls, layer caching, and volume mounts, and will make the server feel slow even under light load.
When possible, put Docker's data directory (/var/lib/docker) on a separate volume from the OS partition. This keeps container churn from competing with the OS for I/O and makes it easier to grow or replace storage independently.
Memory
Reserve 2–4 GB of RAM for the host operating system and the container runtime itself; the remainder is available for running tasks.
Swap (sized roughly equal to physical RAM) can help absorb short bursts of memory pressure, but only enable it on SSD-backed storage — swapping to a spinning disk will tank performance.
Network
Make sure the server has enough bandwidth to pull container images quickly, especially when building or refreshing environments.
If CoderFlow connects to external servers or services — for example, an IBM i system — host it on the same network and as physically close to those systems as possible. Round-trip latency between CoderFlow and the systems it works with has a noticeable impact on source import, build, and debug workflows.