How can I use kerberos under python elasticsearch?

I checked the documentation of elasticsearch:7.x but still didn't find the way how I can force the python ES package to use my kerberos token.

I can reach the ES through a browser, curl ( curl --negotiate -u ) and requests with requests_kerberos ( with mutual OPTIONAL config ) but with the python elasticsearch package calling a simple es.cat.health() I get 401 AuthenticationException...

My questions are:

• where was I wrong, what did I miss?
• is it even possible to force the ES python lib to use kerberos or does it work under the hood?

Thanks in advance.

It seems We've found a solution, but basically not from the docs.

I've read in the doc about the Elasticsearch object can receive http_auth input, but unfortunately I didn't find what it exactly waits for.

So, you can force the Elasticsearch to use the HTTPKerberosAuth class from the requests_kerberos lib in same way as you can use that for a simple get request >

es = Elasticsearch(<host>, http_auth=HTTPKerberosAuth(mutual_authentication=OPTIONAL))

It works well, but any other suggestions, knowledge share or comments are welcomed.