Proxy Configuration

Mark Bakker edited this page Dec 9, 2016 · 1 revision

Why Use a Proxy

If your network configuration restricted outbound traffic, you can proxy all agent traffic through one or several hosts that have more permissive outbound policies.

You have a few options to send traffic to StackState over SSL/TLS for hosts that are not directly connected to the internet.

  1. Using the agent as a proxy (for up to 16 agents per proxy)
  2. Using a web proxy (e.g. Squid, Microsoft Web Proxy) that is already deployed in your network
  3. Using HAProxy (if you want to proxy more than 16-20 agents through the same proxy)

Step-by-step Guides

Using the Agent as a Proxy

  1. Designate one node that runs stackstate-agent as the proxy. In this example assume that the proxy name is proxy-node. This node must be able to reach StackState on the location where you installed StackState.

  2. Allow non-local traffic on proxy-node by changing the following line in /etc/sts-agent/stackstate.conf. # non_local_traffic: no should read non_local_traffic: yes.

  3. Make sure proxy-node can be reached from the other nodes over port 17123. Start the agent on the proxy-node and run on the other nodes:

    curl -v http://proxy-node:17123/status 2>&1 | grep "200 OK"

  4. Update non-proxy nodes to forward to proxy-node. Simply change the following line in /etc/sts-agent/stackstate.conf from:

    dd_url: http://localhost:7070/stsAgent/ to dd_url: http://proxy-node:17123

Using a Web Proxy as Proxy

Traditional web proxies are supported natively by the agent. Simply edit stackstate.conf with your proxy information.

# If you need a proxy to connect to the Internet, provide the settings here
proxy_port: 3128
proxy_user: my_user
proxy_password: my_password

Do not forget to restart the agent for the new settings to take effect.

Using HAProxy as a Proxy

HAProxy is a free, very fast and reliable solution offering proxying for TCP and HTTP applications. While HAProxy is usually used as a load balancer to distribute incoming requests to pools servers, you can also use it to proxy agent traffic to StackState from hosts that have no outside connectivity.

This is the best option if you do not have a web proxy readily available in your network and you wish to proxy a large number of agents. In some cases a single HAProxy instance will be sufficient to handle local agent traffic in your network - each proxy can accommodate upwards of 1000 agents (be aware that this figure is a conservative estimate based on the performance of m3.xl instances specifically. Numerous network-related variables can influence load on proxies. As always, deploy under a watchful eye. Visit for additional information).

agent ---> haproxy ---> StackState

We assume that HAProxy is installed on a host that has connectivity to StackState. You can use the following configuration file if you do not already have it configured.

# Basic configuration
    log local0
    maxconn 4096
    stats socket /tmp/haproxy

# Some sane defaults
    log     global
    option  dontlognull
    retries 3
    option  redispatch
    timeout client 5s
    timeout server 5s
    timeout connect 5s

# This declares a view into HAProxy statistics, on port 3835
# You do not need credentials to view this page and you can
# turn it off once you are done with setup.
listen stats :3835
    mode http
    stats enable
    stats uri /

# This declares the endpoint where your agents will connect.
# In this example we use port 3834 but you can use any other
# free port.
frontend forwarder
    bind *:3834 # DTDG
    mode tcp
    default_backend stackstate

# This is the StackState server. In effect any TCP request coming
# to the forwarder frontend defined above will be proxied to
# StackState's public endpoints.
backend stackstate
    balance roundrobin
    mode tcp
    option tcplog
    server mothership check port 80

Once the HAProxy configuration is in place, you can reload it or restart HAProxy.

We recommend having a cron job that reloads HAProxy every 10 minutes (usually doing something like service haproxy reload) to force a refresh of HAProxy's DNS cache, in case StackState fails over to another IP.

Then edit each agent to point to HAProxy by setting its dd_url to the address of HAProxy (e.g. This dd_url setting can be found in stackstate.conf.


Before you restart the agent you will need to edit your supervisor configuration to disable SSL certificate verification. This is needed to prevent python from complaining about the discrepancy between the hostname on the SSL certificate and your HAProxy hostname.

On GNU/Linux:

You need to edit the supervisor configuration found at:

  • /etc/sts-agent/supervisor_stsagent.conf on debian-based systems
  • /etc/sts-agent/supervisor.conf on redhat-based systems

Assuming that file is found at SUP_FILE

sed -i 's/ --sslcheck=0/' SUP_FILE

On Windows (Starting from agent 3.9.2):

You need to edit your configuration file (stackstate.conf) and add this option:

skip_ssl_validation: yes

Finally restart the agent.

To verify that everything is working properly, you can review the HAProxy statistics at

You can’t perform that action at this time.
You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.
Press h to open a hovercard with more details.