Helm Debt Analysis
Hello ! Here you will find a tool that's hopefully useful! You can reach out to me:
- LinkedIn: berkulsoy
- Twitter: berkulsoy 🇬🇧
- Twitter: berkulsoy_tr 🇹🇷
Those who run kubernetes platforms almost always leverage software from CNCF ecosystem.
Just like our own stuff, those third party dependencies keep getting new releases.
Unseparable part of kubernetes maintenance lifecycle is about keeping up with all those.
What does keeping it up to date means when it comes to CNCF ecosystem ?
Let's see some items:
- Regularly check if a thirdparty dependency has a new version
- See if that new version brings new functionalities, security or bug fixes
- Check the requirements, steps of upgrades. Anything deprecated ? Changed ? Config still usable? Anything potentially breaks?
- Bring it to your test rig to see if it really installs, upgrade goes fine, passes your quality gates etc
- Check if your other stuff that depends on this piece, still works ok
- If all lights are green, propagete it to higher environments
- Doing this cycle for every thirdparty dependency
But what kind of effort should we expect ? How frequently this can happen ?
Typically, the effort correlates with the amount of releases that we have to track. However, keeping upgrade cycle too long, often makes it more painful due to increased code difference integration compatibility issues.
So it's usually better to update often. But how many times do you get updates for your stack ?
If you are curious, write your 3rd party dependency helm charts with repository/chartname format line by line and hit!
Thanks to Artifacthub, I will try to show you what it means.