Kapacitor scraping requests send the authorization field in an unauthenticated request

With the following Kapacitor scraping configuration we are seeing the remote endpoint return `401 Unauthorized`:
```
[[scraper]]
enabled = true
name = "prometheus-test"
discoverer-id = "mystatic"
discoverer-service = "static-discovery"
db = "prometheus_raw"
rp = "autogen"
type = "prometheus"
scheme = "http"
metrics-path = "/api/v1/nodes/knode1/proxy/metrics/cadvisor"
scrape-interval = "30s"
scrape-timeout = "30s"

[[static-discovery]]
enabled = true
id = "mystatic"
targets = ["localhost:8001"]
```

This is due to the fact that even though we have configured no authentication in our scraping configuration, Kapacitor will still send `Authorization: Basic Og==` in the header of the scraping request. This causes the endpoint to try and validate the credentials, even though sending no credentials will result in a successful request (manually tested stripping the Authorization header field with success).

HTTP Request and Response:
```
GET /api/v1/nodes/jng4knode1/proxy/metrics/cadvisor HTTP/1.1
Host: localhost:8001
User-Agent: Prometheus/
Accept: application/vnd.google.protobuf;proto=io.prometheus.client.MetricFamily;encoding=delimited;q=0.7,text/plain;version=0.0.4;q=0.3,*/*;q=0.1
Authorization: Basic Og==
X-Prometheus-Scrape-Timeout-Seconds: 30.000000
Accept-Encoding: gzip
Connection: close

HTTP/1.1 401 Unauthorized
Content-Length: 129
Content-Type: application/json
Date: Tue, 22 Oct 2019 23:04:20 GMT
Connection: close

{"kind":"Status","apiVersion":"v1","metadata":{},"status":"Failure","message":"Unauthorized","reason":"Unauthorized","code":401}
```

It seems like we should omit the authorization header field if the configuration has not defined any authentication.

Kapacitor version:
```
root@kapacitor-b24zw:/# kapacitor version
Kapacitor OSS 1.5.2 (git: HEAD 3086452d00830e01d932838d8c6d1df818648ad3)
```

Installed from `https://dl.influxdata.com/kapacitor/releases/kapacitor_1.5.2_amd64.deb`, and running on 64-bit Ubuntu 16.04

No source file, entry point, or test is named. Reproduce the shown scraper configuration, then trace how scraping requests construct authentication headers; done means an unauthenticated configuration sends no Authorization header while configured credentials still work, with a regression test covering both cases.

Written by the indexing model from the issue text.