The AppDynamics proxy is a Java daemon process that handles communication between the following agents and the Controller:
- Web Server
The proxy reports data collected by the agent to the Controller, which stores, baselines, and analyzes it.
This topic presents basic information to help you examine and resolve proxy issues that may prevent an application agent from connecting to the Controller or reporting data correctly. For detailed information about how the proxy works with a specific agent, see the appropriate Related pages.
Single or Multi-Tenant
For the Node.js and PHP agents, the proxy can be single-tenant (one proxy per agent) or multi-tenant (multiple agents communicating through a single proxy), depending on the way the agents are configured. The default setup for these agents is single-tenant, but there are circumstances for which a mutli-tenant proxy is required or desirable. The Apache Web Server and Python agents always communicate through a multi-tenant proxy when there are multiple agents on a machine.
Automatic or Manual Startup and Shutdown
Typically the proxy is automatically started when the application agent starts up, but in some cases, you need to launch it manually. These cases vary depending on both the particular agent and the application environment.
After the proxy is started (automatically or manually) it registers with the Controller and requests the agent configuration. The agent must receive the configuration from the Controller via the proxy before it can report metrics.
If the proxy was automatically started, the agent is supposed to shut down the proxy as part of its own cleanup procedures. If the proxy was manually started, it must be shut down manually.
If the proxy is running, you can check the proxy log to examine connection issues.
The proxy log files are named
If the proxy is not running, examine the agent logs for the typical Java startup signature to see if the proxy started.
See the agent-specific documentation for the location of these logs:
- Node.js: Default is
/tmp/appd/logs.See Install the Node.js Agent for details
- PHP: Default is
<php_agent_install_dir>/logs.See Install the PHP Agent for more details
- Python: Default is
/tmp/appd/logs.See Python Agent Debugging and Logging for details
- Web Server:
Proxy Not Started Issues
The agent will not work if the proxy did not start.
Determine whether the proxy started by running this command:
If the proxy is running, you should see
proxy in the output, something like this:
If the proxy did not start, the most common reason is insufficient permissions.
The agent installation directory, and its proxy control subdirectory, must be readable and executable by all and writable by the directory owner:
The agent installation directory must be owned by the instrumented application user. Who this user is depends on the platform. It could be the Apache user, the python container user, the nginx user, the
If the proxy did not start, set these permissions and try again.
Proxy Connection Issues
If the proxy started but you don't have a connection to the Controller, you may be using the wrong controller information.
Examine the proxy log and verify that there are no typos in the Controller domain name or port and that SSL setting is enabled for an SSL connection. If there are mistakes, edit the Controller/port/SSL values in the configuration. These settings are on the first page of the Agent Download and Install Wizard or the agent settings (require statement for Node.js) if you installed the agent manually.
If you have verified that the Controller settings are correct and the proxy still does not connect to the Controller, telnet the SaaS Controller:
If you cannot access the Controller via telnet, examine the agent-registered messages in the logs that indicate why the proxy is unable to connect, such as the existence of a firewall or other obstacle at your site. Work with your administrator to see if you can resolve the issue.
If you can reach the Controller via telnet, there may be a problem on the Controller side.