[9.1](backport #45092) [Elasticsearch] Allow Env Injected Auth

[mergify]

This allows tools to inject the auth parameters via the environment, which enables usage of dynamically deciding what to use for auth in OTel mode while still supporting blank values.

Specifically, this allows end users to specify either username and password or the API Key via configuration, which is not currently possible in the OTel configuration due to #45102.

Proposed commit message

Allow ELASTICSEARCH_READ_USERNAME, ELASTICSEARCH_READ_PASSWORD, and ELASTICSEARCH_READ_API_KEY to be specified in the environment to control auth for Elasticsearch in a receiver context (not the exporter / output layer of Elasticsearch). This also enables the usage of them for the autoops_es module.

Disruptive User Impact

I intentionally added READ_ to avoid disruption. These variables also never existed before, so anyone using them would have been using them to set the configuration already, which helpfully avoids collisions.

Author's Checklist

• These default to blank, so no functionality really changes.

How to test this PR locally

• Try using the Elasticsearch module with and without these environment variables.
• Try using the Elasticsearch module with environment variables, but also config that overrides them and observe that they are overridden (easiest to provide invalid auth in env, but valid in config to test).
• Try setting all three at the same time (Metricbeat will fail to start because that is an invalid setup just like putting all three directly in config)

Related issues

• Relates [OTel] Unset / blank environment variables are misinterpreted as objects #45102

Use cases

• Use elasticsearch and autoops_es modules via Kubernetes / Docker
• Use those modules with something dynamically specifying the auth (i.e., Buildkite) without three (no auth, user/pass, and api key) separate config files

  ---

  This is an automatic backport of pull request [Elasticsearch] Allow Env Injected Auth #45092 done by Mergify.

[github-actions]

🤖 GitHub comments

Expand to view the GitHub comments

Just comment with:

• run docs-build : Re-trigger the docs validation. (use unformatted text in the comment!)