pushprox

README

PushProx

PushProx is a client and proxy that allows transversing of NAT and other similar network topologies by Prometheus, while still following the pull model.

While this is reasonably robust in practice, this is a work in progress.

Running

First build the proxy and client:

go get github.com/e-bits/pushprox/{client,proxy}
cd ${GOPATH-$HOME/go}/src/github.com/e-bits/pushprox/client
go build
cd ${GOPATH-$HOME/go}/src/github.com/e-bits/pushprox/proxy
go build

Run the proxy somewhere both Prometheus and the clients can get to:

./proxy

On every target machine run the client, pointing it at the proxy:

./client --proxy-url=http://proxy:8080 --fqdn=client:9100

In Prometheus, use the proxy as a proxy_url:

scrape_configs:
- job_name: node
  proxy_url: http://proxy:8080/
  static_configs:
    - targets: ['client:9100']  # Presuming the FQDN of the client is "client".

If the target must be scraped over SSL/TLS, add:

  params:
    _scheme: [https]

rather than the usual scheme: https. Only the default scheme: http works with the proxy, so this workaround is required.

Service Discovery

The /clients endpoint will return a list of all registered clients in the format used by file_sd_configs. You could use wget in a cronjob to put it somewhere file_sd_configs can read and then then relabel as needed.

scrape_configs:
- job_name: node
  proxy_url: http://proxy:8080/
  file_sd_configs:
    - files:
      - pushproxclients.json

How It Works

The client registers with the proxy, and awaits instructions.

When Prometheus performs a scrape via the proxy, the proxy finds the relevant client and tells it what to scrape. The client performs the scrape, sends it back to the proxy which passes it back to Prometheus.

Security

There is no authentication or authorisation included, a reverse proxy can be put in front though to add these.

Running the client allows those with access to the proxy or the client to access all network services on the machine hosting the client.