Failover Protection with Traffic Distributor

Traffic Distributor includes a health-check module that regularly verifies backend availability and automatically removes unavailable servers from request routing.

The health-check functionality is enabled by default. Its interval, failure threshold, recovery threshold, timeout, initial backend state, port, and protocol can be adjusted when the default behaviour does not match the application requirements.

On this page

Open the NGINX Configuration
Health-Check Configuration Syntax
Health-Check Parameters
Default Failover Example
Apply Changes without Downtime
What’s next?

Open the NGINX Configuration

1Open the NGINX Config panel

In the Traffic Distributor environment, click the Config button for the NGINX layer.

Navigate through Root > etc > nginx and open:

/etc/nginx/nginx-jelastic.conf

Traffic Distributor NGINX configuration files — Open nginx-jelastic.conf from the Traffic Distributor NGINX configuration panel.

Health-Check Configuration Syntax

Scroll to approximately line 50 and locate the health-check directive inside the upstream common section.

check interval={interval} fall={fail_count} rise={rise_count}
      [timeout={timeout}] [default_down={true/false}]
      [port={port}] [type={type}]

Traffic Distributor failover parameters — The health-check directive inside the upstream common section.

Parameters shown in square brackets are optional. Required values define the check interval and the failure and recovery thresholds.

Health-Check Parameters

Parameter	Description
`interval`	Delay in milliseconds between two consecutive health-check requests.
`fall`	Number of consecutive failed checks after which the backend is marked unavailable.
`rise`	Number of consecutive successful checks after which the backend is marked available again.
`timeout`	Maximum time in milliseconds that the module waits for a backend response before treating the check as failed.
`default_down`	Sets the initial backend state before the required checks complete. `true` starts it as down; `false` starts it as up. The documented default is `true`.
`port`	Port used for the check. A value of `0` uses the default server port for the selected protocol.
`type`	Protocol used to verify backend availability.

Supported Health-Check Types

tcp — opens a TCP socket connection.
ssl_hello — sends an SSL Client Hello and expects a Server Hello response.
http — sends an HTTP request and parses the response.
mysql — connects to a MySQL server and receives its greeting response.
ajp — sends an AJP Cping packet and expects a Cpong response.
fastcgi — sends a FastCGI request and parses the response.

Default Failover Example

The documented example checks both backends every three seconds and expects a normal HTTP response with status code 200.

Each backend is checked every 3 seconds.
After 3 consecutive failed checks, the backend is marked down.
The unavailable backend is removed from routing.
All requests are sent to the remaining environment.
After 3 successful checks in a row, the recovered backend is added back automatically.

Traffic continuity: The second backend must be healthy and capable of handling the complete workload when the first backend is excluded.

Apply Changes without Downtime

2Save the configuration

Save the updated nginx-jelastic.conf file.

3Reload Traffic Distributor

Open the Traffic Distributor add-on menu and select Reload configuration.

Reload Traffic Distributor configuration — Apply the updated failover settings through a graceful NGINX reload.

Confirm the action in the pop-up window. The graceful reload applies the new failover settings without restarting the complete server, helping avoid application downtime.

Before production changes, verify the NGINX syntax and confirm that the selected path, port, protocol, timeout, and thresholds match the real backend behaviour.