cannot read file nginxlog-exporter.conf

[scottie1996]

Issue Summary

Title: Prometheus Nginxlog Exporter: Config File Unsupported File Type Error

Description:

I encountered an issue while setting up the prometheus-nginxlog-exporter. Despite following the instructions to create a configuration file, the exporter fails to start, citing an "unsupported file type" error. Below are the detailed steps, error messages, and attempted resolutions.

Steps to Reproduce:

1. Environment Details:

   • OS: CentOS 7
   • Prometheus Nginxlog Exporter version: 1.11.0

2. Initial Configuration:

   • Created a configuration file /opt/nginxlog-exporter.conf in YAML format:

     log_file: /var/log/nginx/access.log
     metrics:
     - name: http_requests_total
      match: 'GET /'
      labels:
        method: GET
     - name: http_requests_total
      match: 'POST /'
      labels:
        method: POST

3. Service Configuration:

   • Created a systemd service file /etc/systemd/system/nginxlog-exporter.service:

     [Unit]
     Description=Prometheus Nginxlog Exporter
     After=network.target
     
     [Service]
     ExecStart=/opt/prometheus-nginxlog-exporter -config-file /opt/nginxlog-exporter.conf
     Restart=always
     
     [Install]
     WantedBy=multi-user.target

4. Errors Encountered:

   • When starting the service, the following error is logged:

     config file '/opt/nginxlog-exporter.conf' has unsupported file type

Error Logs:

2024-06-14T17:33:07.870+0800 fatal prometheus-nginxlog-exporter/main.go:164 config file '/opt/nginxlog-exporter.conf' has unsupported file type
main.loadConfig
/home/runner/work/prometheus-nginxlog-exporter/prometheus-nginxlog-exporter/main.go:164
main.main
/home/runner/work/prometheus-nginxlog-exporter/prometheus-nginxlog-exporter/main.go:117
runtime.main
/opt/hostedtoolcache/go/1.18.10/x64/src/runtime/proc.go:250

Resolution Steps Taken:

1. Identified Issue with File Format:

   • Realized that the prometheus-nginxlog-exporter expects an HCL (HashiCorp Configuration Language) formatted configuration file.

2. Updated Configuration File to HCL Format:

   • Modified /opt/nginxlog-exporter.conf to the following HCL format:

     listen {
     port = 9113
     }
     
     namespace "nginx" {
     format = "$remote_addr - $remote_user [$time_local] \"$request\" $status $body_bytes_sent \"$http_referer\" \"$http_user_agent\""
     
     source {
      files = ["/var/log/nginx/access.log"]
     }
     
     labels {
      app = "nginx"
     }
     
     counters {
      request_count {
        help = "Number of HTTP requests"
        match = ".*"
        name = "http_requests_total"
      }
     }
     
     histograms {
      request_duration {
        help = "Histogram of request duration"
        match = ".*"
        name = "http_request_duration_seconds"
        buckets = [0.005, 0.01, 0.025, 0.05, 0.1, 0.25, 0.5, 1, 2.5, 5, 10]
      }
     }
     }

3. Manual Testing:

   • Ran the exporter manually to verify configuration:

     /opt/prometheus-nginxlog-exporter -config-file /opt/nginxlog-exporter.conf

   • This command worked without errors.

4. Restarted Service:

   • Restarted the systemd service:

     sudo systemctl start nginxlog-exporter
     sudo systemctl status nginxlog-exporter

5. Verified Prometheus Target:

   • Checked Prometheus targets at http://10.0.89.39:9090/targets and confirmed that the nginxlog-exporter is now running and reporting metrics.

Conclusion:

The issue was caused by the incorrect file format for the configuration file. The exporter expects HCL format, not YAML. Updating the configuration file to HCL format resolved the issue.