containers icon indicating copy to clipboard operation
containers copied to clipboard

Establish collection of health checks

Open yarikoptic opened this issue 1 year ago • 1 comments

In the light of current xz incident, I think it would be great to provide dinner kind of health status monitoring across containers we contain. And run it not only when we add a container but correspondingly when we add a new check. So in some ways similar to https://github.com/dandi/dandisets-healthstatus?tab=readme-ov-file in that we want

  • establish collection of tests with each test versioned (eg output'ing --version)
  • specification of tests to run for a file type (regex )
  • per each file/version contain a list of tests and versions of the tests

We will need to run it where we have container images but such a tool sounds generic and could run as CI as well .

@jwodder do you may be know already existing framework? If not, it is for us to create one I think. Now I am thinking even may be dandi healthstatus potentially could be generalized into it eventually.

yarikoptic avatar Apr 08 '24 08:04 yarikoptic

@yarikoptic

do you may be know already existing framework?

No.

Now I am thinking even may be dandi healthstatus potentially could be generalized into it eventually.

I don't think so. Much of dandisets-healthstatus deals with mounting & traversing file hierarchies to run tests on individual files, which doesn't seem to apply here.

jwodder avatar May 20 '24 19:05 jwodder