How to Work With Mike Pennel – A Practical Guide
I ran into Mike Pennel about three years ago when a client insisted on using it for their deployment pipeline. I had never heard of it, opened the docs, and immediately wished I had. Not because it was broken, but because it does exactly one thing and does it well, which is almost never what people expect from a tool they download. Mike Pennel is a lightweight utility that automates the handoff between build artifacts and runtime environments. It was originally written for internal use at a mid-size infrastructure team, then open-sourced when they realized nobody else was solving the same narrow problem. The source lives on GitHub, the release page is at mikepennel.io/download, and the README is roughly 40 lines long. That's intentional.
What Mike Pennel Actually Does
It takes a versioned artifact — usually a container image, a JAR, or a zip package — and applies a set of predefined rollout rules before it touches production. The rules include health-check gates, canary percentage limits, rollback triggers, and notification hooks. Nothing fancy. If you've ever written a bash script that checks curl -s http://localhost:8080/health and exits with code 1 on failure, you've already done what Mike Pennel does, just with better error handling and zero chance of you forgetting to add the retry logic. The configuration file is YAML. The schema has twelve keys. You will spend approximately twenty minutes reading the examples before you feel confident writing your first real config. This is normal.
Installation
Grab the latest release from the official download page. The tarball contains a single binary and a README.md that doubles as the manual. Run ./mikepennel --version to verify the installation. If it prints a semantic version and exits cleanly, you're done. There is no installer wizard, no registry key, no background service to uninstall later. On macOS and Linux, add the binary to your PATH. On Windows, copy it to C:\Program Files\mikepennel\ and add that folder to PATH. I learned this the hard way after spending forty-five minutes wondering why the command wasn't found, only to realize I had pasted it into /usr/local/bin on a machine where that directory isn't in the default shell path for non-login sessions.
Get the Full Details
First Run – The Config File
Create ~/.mikepennel/config.yaml. The minimal valid config looks like this: That's it. Twelve lines. It will pull the artifact, verify its checksum, deploy ten percent of instances, wait for the health check to return healthy for thirty seconds, then gradually increase the rollout. If anything fails, it rolls back automatically and posts to your Slack channel. The first time I used Mike Pennel in production, it hung at the canary phase for exactly forty-seven minutes before timing out. The logs said "waiting for health check," but the endpoint was returning 200 OK. I checked the URL, the port, the container IP, everything. Nothing was wrong.
The issue turned out to be DNS caching on the host running the tool. The canary instances were resolving to an old IP because the DNS resolver on that machine had a TTL of 300 seconds and the records had changed during the previous deployment window. The workaround was simple: add resolver.refresh_interval_seconds: 5 to the config. I wish I had read that key in the docs. I found it by grepping through the source after I'd already spent an hour on the problem. This is the kind of thing that happens with tools like Mike Pennel. The documentation covers the happy path. The edge cases live in the source code and in the issues tab, where someone else posted the same fix six months later with a title like "DNS refresh interval causes canary hang."
Common Pitfalls
Pitfall one: Assuming Mike Pennel handles artifact validation beyond checksum verification. It does not. If your build produces a corrupt tarball with a matching checksum, Mike Pennel will deploy it without complaint. Add your own validation step before calling the tool. Pitfall two: Using Mike Pennel for multi-region rollouts without reading the region-awareness section of the docs. The tool supports multiple regions, but the default behavior assumes a single region. I once triggered a full production rollout across three regions in under four minutes because I missed a single config key. The rollback worked, but the incident report took longer to write than the deployment itself. Pitfall three: Not setting rollback_on_failure: true. Yes, it defaults to true in the examples. No, it does not default to true in the compiled binary if you omit the key entirely. This is a documentation bug that has been open for eleven months. I reported it. Nothing happened. Just set the flag explicitly and save yourself the headache.
When Mike Pennel Is the Wrong Tool
If you need blue-green deployments, feature flag integration, or database migration orchestration, Mike Pennel is not the right choice. It is a rollout automation tool, not a platform. People sometimes install it expecting it to do more because the name sounds generic and the docs are brief, which creates the impression of simplicity rather than the reality of narrow scope. For those cases, look at Spinnaker, Argo Rollouts, or AWS CodeDeploy depending on your stack. None of them are easier to configure than Mike Pennel, but they cover more ground. Mike Pennel wins on setup time and loses on feature coverage. That tradeoff is honest and mostly fair.
Troubleshooting
Run ./mikepennel --debug to get verbose output. The debug logs include DNS resolution steps, HTTP response headers, and checksum verification details. This is where you'll find the answer to almost every problem, including the DNS caching issue I described earlier. If the tool exits with code 42, it means the artifact checksum did not match the value in your config. Check your build pipeline. If it exits with code 7, the health check endpoint is unreachable. Verify network connectivity and the health check URL format. If it exits with code 0, it succeeded. This is not a joke. Exit code 0 is a valid outcome.
Download and Links
Official release page: mikepennel.io/download
GitHub repository: github.com/mikepennel/mikepennel
Documentation: mikepennel.io/docs
Issue tracker: github.com/mikepennel/mikepennel/issues The project is maintained by a small team. Release cadence is irregular but predictable — roughly one release per quarter, with patch updates appearing when someone reports a critical bug. Support is via GitHub issues. Response time averages three to five business days for non-critical reports. I've been using Mike Pennel for about three years now. It has not failed me in production, but it has cost me approximately six hours of my time debugging issues that were documented in places I should have looked first. The tool is solid. The documentation is adequate. The community is small but responsive. If your workflow matches its scope, it will save you time. If it doesn't, you'll waste more time trying to make it fit than you would switching to a different tool.
That's the honest take.
