Systems Projects Level 5-6¶

These projects work close to the operating system and infrastructure — the kind of tooling that keeps systems running and observable. They exercise the Embedded & Systems and Performance material.

1. System monitor + alerter¶

Goal: collect CPU/memory/disk metrics, alert when thresholds are crossed.

• Skills: psutil, threshold logic, notifications, running as a service.
• Minimal version: print metrics on a loop, alert when CPU > 90%.
• Stretch: export to Prometheus, a dashboard, consecutive-breach logic, run as a systemd service. Directly builds the System Monitoring example out into a real tool.

2. Log analyzer / aggregator¶

Goal: parse log files, extract structured data, and report patterns (error rates, top offenders).

• Skills: file I/O, regex, collections, maybe a small web view.
• Minimal version: count log levels (ERROR/WARN/INFO) in a file.
• Stretch: tail live logs, time-window analysis, alert on error spikes, handle multiple formats. Great use of Automation & Scripting.

3. A profiler / benchmark harness¶

Goal: a tool that times and compares implementations, reporting stats.

• Skills: time.perf_counter, statistics, decorators, reporting.
• Minimal version: a decorator that times a function and prints the result.
• Stretch: multiple runs with percentiles, comparison tables, memory tracking, plots. Applies the Profiling topic.

4. A CLI dev-ops tool¶

Goal: a polished command-line tool that automates a real workflow (deploys, backups, environment setup).

• Skills: argparse/Click/Typer, subprocess, config files, good UX and error messages.
• Minimal version: one subcommand that does something useful with clear --help.
• Stretch: multiple subcommands, config-driven, dry-run mode, packaged for pip install. See CLI Tools and Packaging.

5. Config-drift detector¶

Goal: compare a system's actual config against a desired spec and report differences.

• Skills: parsing config, set/diff logic, idempotency thinking.
• Minimal version: diff two config files and print add/remove.
• Stretch: apply changes, multiple hosts, git-backed desired state. This generalizes the tested diff from Networking Automation.

6. A tiny service supervisor¶

Goal: start, monitor, and restart child processes if they die (a mini process manager).

• Skills: subprocess, signals, health checks, graceful shutdown.
• Minimal version: launch a process and restart it if it exits.
• Stretch: multiple processes, config file, restart backoff, log capture. Teaches the mechanics behind systemd/supervisord.

Why systems projects are valuable¶

They force you to deal with the messy real world: processes, signals, files, the OS, failure. This "ops-aware" skill set — building tools that run unattended, observe systems, and recover from failure — is exactly what distinguishes a developer who can operate software from one who can only write it.

For the frontier, see Research Projects.