Elastic-Agent snapshot downloader ignores proxy settings

[michel-laterman]

Elastic-Agent (running under fleet) does not use any proxies settings when looking up snapshot locations for artifact upgrade.
Snapshot lookup uses the standard library client (gohttp is an alias for net/http)

beats/x-pack/elastic-agent/pkg/artifact/download/snapshot/downloader.go

Lines 55 to 56 in 6af8ffb

	artifactsURI := fmt.Sprintf("https://artifacts-api.elastic.co/v1/search/%s-SNAPSHOT/elastic-agent", version)
	resp, err := gohttp.Get(artifactsURI)

This ignores any settings defined in fleet.yml, specifically:

agent:
  download:
    proxy_url: PROXY_ADDRESS

If the agent is running behind an Internet proxies the logs will show that timeouts on the search request occured:

{"log.level":"error","@timestamp":"2021-09-13T20:28:18.431Z","log.origin":{"file.name":"log/reporter.go","file.line":36},"message":"2021-09-13T20:28:18Z - message: Application: [59f55f0a-93bf-4462-ba7f-3de75161046a]: State changed to FAILED: initiating verifier: failed to detect remote snapshot repo, proceeding with configured: Get \"https://artifacts-api.elastic.co/v1/search/7.15.0-SNAPSHOT/elastic-agent\": dial tcp 35.188.12.98:443: i/o timeout - type: 'ERROR' - sub_type: 'FAILED'","ecs.version":"1.6.0"}

For confirmed bugs, please report:

• Version: 7.14.1
• Steps to Reproduce:

1. Deploy ES + Kibana v7.15.0-SNAPSHOT
2. Deploy fleet/elastic-agent v7.14.1 behind an Internet proxy
3. Attempt to use the Fleet UI to update the running process to v7.15.0-SNAPSHOT

[elasticmachine]

Pinging @elastic/agent (Team:Agent)