reverse_exporter.complete.yml

# This is *every* reverse_exporter config option in one-file.
# This file should not normally used this way - specific only what you need.

web:
  # timeout to receive headers from connections - prevent Slow Loris
  read_header_timeout: 1s
  # list of addresses to listen on
  listen:
    # open a Unix socket file on /var/run/server
    - unix:///var/run/server.socket
    # listen on tcp port 9998.
    - tcp://0.0.0.0:9998
    # listen on port 9998 with TLS
    - tcps://0.0.0.0:9998?tlscert=/path/to/file/in/pem/format.crt&tlskey=/path/to/file/in/pem/format.pem
    # listen on Unix socket file on /var/run/server with TLS
    - unixs:///var/run/server.socket?tlscert=/path/to/file/in/pem/format.crt&tlskey=/path/to/file/in/pem/format.pem
    # listen on 9998 with TLS and TLS client auth
    - tcps://0.0.0.0:9998?tlscert=/path/to/file/in/pem/format.crt&tlskey=/path/to/file/in/pem/format.pem&tlsclientca=/path/to/cert

# Each item in the list is the name of a url subpath to combine exporters under.
reverse_exporters:
# the normal use of this exporter is intended to be presenting a consistent
# /metrics endpoint for appliance-like environments which may contain multiple
# exporters.
- path: /metrics
  # auth_type configures password protection on the endpoint
  auth:
    basic_auth:
      - username: root
        password: test
  exporters:
    http:
    - name: prometheus
      address: http://127.0.0.1:9090/metrics
      # proxy_timeout sets the TCP deadline for a request to complete. No data
      # is returned by an endpoint until all exporters have received data, so keep
      # this low (or preferably: don't use network endpoints).
      # a value of 0 means no timeout (the request can potentially stall until the
      # client terminates, which Prometheus will do by default).
      timeout: 1s
    - address: http://127.0.0.1:9100/metrics
      name: node_exporter
      # labels is a map of additional static labels to add (in addition to the
      # enforced "name" field)
      labels:
        node_uuid: some.special.identifier
    # metrics from jobs inside a container can be easily included provided they are
    # in the text exposition format. Just path a file URI as the address.
    file:
    - name: cron_metrics
      path: example.metrics.prom
    # an executable script can also be passed with a special `exec` URL. exec proxy's have two modes:
    # in non-caching mode, the proxy runs an instance of the script as soon as a request is received,
    # and buffers up additional requests - returning the result data to all of them once execution is
    # finished. Only a single instance is ever executing.
    exec:
    - name: dynamic_metrics
      command: ./scripted_metrics.sh
      args: ["arg1", "arg2"]
    # In caching mode, the command is executed continuously with a given timeout, and cached results
    # are served to Prometheus instances. Your script should probably include a timestamp in this mode.
    exec_cached:
    - name: periodic_dynamic_metrics
      command: ./scripted_metrics.sh
      args: []
      # interval to execute the script over
      exec_interval: 30s

# The exporter does support declaring arbitrary paths, for example if you were
# fronting something like the blackbox_exporter which changes its return based
# on the Prometheus query string.
- path: /blackbox
  exporters:
    http:
    - name: blackbox_exporter
      address: http://127.0.0.1:9998/probe
      # disable appending the name (above) to the exporter. Important: if you end
      # up combining multiple metrics with undistinguished names, the endpoint
      # will disable itself.
      no_rewrite: true
      # ForwardURLParams determines whether the exporter will have ALL url params
      # of the parent request added to it.
      forward_url_params: true