Relevant telegraf.conf:
# contents does not matter, as it is never read:
[[inputs.http_response]]
urls = [ "https://www.influxdata.com/" ]
response_timeout = "5s"
follow_redirects = false
interval = "120s"
System info:
(most likely OS independent)
RHEL7, RHEL8, ...
Steps to reproduce:
- Create the above configuration in the file https::##www.influxdata.com.conf:
cat > 'https:##www.influxdata.com'.conf << EOF
[[inputs.http_response]]
urls = [ "https://www.influxdata.com/" ]
response_timeout = "5s"
follow_redirects = false
interval = "120s"
EOF
- Create a "verification" link:
ln 'https:##www.influxdata.com'.conf www.influxdata.com.conf
- Make telegraf fail on the first:
telegraf --test --config 'https:##www.influxdata.com'.conf
2021-04-29T12:41:46Z I! Starting Telegraf 1.18.0
2021-04-29T12:41:46Z E! [telegraf] Error running agent: Error loading config file https:##www.influxdata.com.conf: Retry 0 of 3 failed connecting to HTTP config server Get "https:#%23www.influxdata.com.conf": http: no Host in request URL
- and succeed on the other:
telegraf --test --config www.influxdata.com.conf
2021-04-29T12:42:34Z I! Starting Telegraf 1.18.0
> http_response,host=...,method=GET,result=success,server=https://www.influxdata.com/,status_code=200 content_length=224068i,http_response_code=200i,response_time=0.397142678,result_code=0i,result_type="success" 1619700155000000000
Expected behavior:
Telegraf should read the configuration file if it exists. The filename is valid and file exists.
Actual behavior:
Telegraf tries to fetch a non-existing URL
Additional info:
We use the naming standard as we perform remote monitoring of thousands of servers.
We use the naming scheme in a lot of other places too (kapacitor being one of them).
Bug is verified present in 1.14.3 and 1.18.0
Relevant telegraf.conf:
System info:
(most likely OS independent)
RHEL7, RHEL8, ...
Steps to reproduce:
Expected behavior:
Telegraf should read the configuration file if it exists. The filename is valid and file exists.
Actual behavior:
Telegraf tries to fetch a non-existing URL
Additional info:
We use the naming standard as we perform remote monitoring of thousands of servers.
We use the naming scheme in a lot of other places too (kapacitor being one of them).
Bug is verified present in 1.14.3 and 1.18.0