This project provisions and operates a public Ubuntu server in Oracle Cloud's Johannesburg region. It automates the network, retries instance creation when capacity is unavailable, and serves this project page from nginx over HTTPS. The goal is a small, reproducible home for experiments, services, and infrastructure work.
[2026-09-22 20:16:45] >>> Attempt 557: launching linux-server [2026-09-22 20:18:24] Launch error: TooManyRequests: Too many requests for the user [2026-09-22 20:18:24] >>> Attempt 557: launch failed. Retrying in 600s [2026-09-22 20:28:30] >>> Attempt 558: launching linux-server [2026-09-22 20:30:07] Launch error: InternalError: Out of host capacity. [2026-09-22 20:30:07] >>> Attempt 558: launch failed. Retrying in 600s [2026-09-22 20:40:13] >>> Attempt 559: launching linux-server [2026-09-22 20:41:50] Launch error: InternalError: Out of host capacity. [2026-09-22 20:41:50] >>> Attempt 559: launch failed. Retrying in 600s [2026-09-22 20:51:55] >>> Attempt 560: launching linux-server [2026-09-22 20:53:32] Launch error: TooManyRequests: Too many requests for the user [2026-09-22 20:53:32] >>> Attempt 560: launch failed. Retrying in 600s [2026-09-22 21:03:37] >>> Attempt 561: launching linux-server [2026-09-22 21:05:14] Launch error: InternalError: Out of host capacity. [2026-09-22 21:05:14] >>> Attempt 561: launch failed. Retrying in 600s [2026-09-22 21:15:19] >>> Attempt 562: launching linux-server [2026-09-22 21:16:56] Launch error: InternalError: Out of host capacity. [2026-09-22 21:16:56] >>> Attempt 562: launch failed. Retrying in 600s
{ "provider": "Oracle Cloud Infrastructure", "region": "af-johannesburg-1", "compute": "VM.Standard.A1.Flex / 4 OCPUs / 24 GB", "network": "VCN / subnet / internet gateway / route table", "delivery": "Ubuntu / nginx / Cloudflare HTTPS" }
The provisioner reuses existing OCI resources when possible, launches the latest compatible Ubuntu image, waits for a running instance, and prints the SSH connection details. The current public page is served by nginx; Coolify was tested and removed after resource pressure on the smaller AMD host made a lightweight static setup the better fit.
An automated GitHub Actions workflow can keep the capture process moving: it serializes retries, stores the provisioning log, and sends an optional webhook notification when the Flex instance is confirmed running. For continuous execution beyond a GitHub runner's time limit, the project also documents an always-on systemd service.