Monitor Your Service Stack from the Terminal — Without a SaaS Subscription

Datadog is $15/host/month. New Relic has a free tier with a 100GB/month data cap. If you're running a small fleet of local or self-hosted services — a gateway, a dashboard, a background worker — you don't need either of those.

Last updated: June 2026 · profitmax-pro on GitHub

The problem it solves

When you're running several interconnected services — a local AI gateway, a dashboard, a workspace service — you end up with a mental checklist. Is port 8642 responding? Did the workspace process restart cleanly? Is the state file stale? You open a browser, run a curl, grep a log. It works, but it's manual every time.

profitmax-pro wraps that checklist into a single command with structured pass/fail output. Run it before a deploy, after a restart, or at the start of a session. One command, clear output.

Install

profitmax-pro is not on public PyPI. Install from source (Python 3.11+ required):

git clone https://github.com/Coding-Dev-Tools/profitmax-pro.git
cd profitmax-pro
pip install .

Confirm it's working:

profitmax --task status

You should see a list of the 7 registered health check tools.

Three commands worth knowing

1. profitmax --task status — list registered checks

Shows which health check tools are registered in the tool registry. Run this after install to confirm everything loaded correctly, or after adding a custom tool to verify registration.

2. profitmax --task heartbeat — run all checks

The main command. Runs all 7 built-in checks in sequence and prints pass/fail output for each. Checks include:

• OpenClaw service (port 18789)
• Paperclip service (port 3100)
• KiloClaw file-based state
• Aggregated service health across the registered stack

The sequential executor prints clear output per check — no noise, no dashboard required.

3. profitmax --task self-improve — LLM improvement suggestions

Sends current health check state to an LLM and asks for infrastructure improvement suggestions. Falls back through DeepSeek → Gemini Flash → Mistral, so it uses whichever API key you already have. If none are configured, skip this command — the other two run fully offline.

Adding a custom health check

The tool registry uses a decorator pattern. To add a check for your own service:

# In a new file under pai/tools/
from pai.tool_registry import tool

@tool
def my_service_check():
    """Check my-service is responding on port 9999."""
    import socket
    try:
        with socket.create_connection(("localhost", 9999), timeout=2):
            return {"status": "pass", "service": "my-service"}
    except OSError:
        return {"status": "fail", "service": "my-service", "error": "port 9999 not reachable"}

Import the file and the executor picks up the check automatically on the next profitmax --task heartbeat run. No configuration file needed.

How it compares

Datadog and New Relic pricing sourced from their public pricing pages, June 2026.

When to use it (and when not to)

profitmax-pro is a good fit if you're running a small local service stack, want a structured heartbeat command rather than a dashboard subscription, and don't need alerting, long-term metrics storage, or distributed tracing. It's essentially a test harness for your infrastructure instead of your code.

It's not a fit for production fleet monitoring at scale. If you need distributed tracing, on-call paging, anomaly detection, or cross-region aggregation, a hosted platform makes more sense. profitmax-pro doesn't attempt to compete there.

Frequently asked questions

Is profitmax-pro on PyPI?

No. Install from source with git clone and pip install .. There is no pip install profitmax-pro command that works — PyPI publishing is not yet available. Python 3.11+ required.

What services does it check out of the box?

The 7 built-in tools cover OpenClaw (port 18789), Paperclip (port 3100), KiloClaw (file-based state), and aggregated service health. If those services aren't part of your stack, the checks will fail — add your own via the @tool decorator instead.

Does it require an LLM API key?

Only for --task self-improve. The heartbeat and status commands run entirely offline. The LLM router tries DeepSeek, then Gemini Flash, then Mistral — use whichever provider you have a key for, or skip that command entirely.

Can I run it on a schedule?

Yes — wrap it in a cron job or a Windows Scheduled Task. Since it prints structured output to stdout and exits with a sensible exit code, it's easy to pipe into a log file or a notification script.