Request for an application health check

[crisbratu]

Hey mixu,

We've been running npm_lazy 1.7.0 for a few months, and it's working great. However once in a while the application needs to be restarted due to errors. Would a healthcheck endpoint be feasible to implement? Or is there an easy way, besides trying to hit it with http requests?

Thanks,
Cris

[mixu]

@crisbratu nice, glad to hear that! You could simply check whether some random package index is fetchable, e.g. curl -vvv localhost:8080/starwars. I'm also open to adding a health check endpoint, as long as it doesn't conflict with existing npm urls, e.g. https://github.com/npm/npm-registry-couchapp/blob/master/registry/rewrites.js

[mrmarcsmith]

This would be a nice feature.
I imagine a good place to start would be a simple call and response.

localhost:8080/health returning {"status":"ok"}

not necessarily making a call to npm but just to make sure the process is alive. In Kubernetes, if a container doesn't respond within a given timeout its restarted. we wouldn't want a health check making an external call every 15 seconds and the container could be unnecessarily restarted if npm take longer than a second or so to respond.

@mixu thoughts?