Q: Can the NetScaler generate an HTTP 503 response?

A: Yes, the NetScaler can generate a HTTP 503 response in the following scenarios:

• On a Load Balancing vServer on the NetScaler on which spillover is configured, once the spillover limit is reached, for the subsequent connections, the NetScaler returns the HTTP 503 response.
• When the setting downstateflush is set to Disabled (Enabled by default) on a vServer, if the vServer becomes disabled or goes down, then the existing connections is not closed immediately, instead, the existing connections are permitted to continue until they time out. If the NetScaler receives any more requests on the existing connections, the NetScaler returns the HTTP 503 response to the client.

The NetScaler has a counter si_tot_svr_busy_err which is incremented if the server is returning error codes greater than 500. You can look for this counter’s value to determine if the server is returning this error or not. The following is an example (ran from shell):

root@ns# /var/nslog/nsconmsg -K newnslog -g si_tot_svr_busy_err -d stats

Displaying current counter value information
Performance Data Record Version 2.0
reltime:mili second between two records Mon Feb 23 17:47:33 2009
Index reltime     counter-value symbol-name&device-no
    0       0                 0 si_tot_svr_busy_err

Additional Resources

HTTP Response status codes beginning with the digit 5 (five) indicate cases in which the server is aware that it has encountered an error or is otherwise incapable of performing the request.

The HTTP status code 503 Service Unavailable indicates that the server is currently unable to handle the request because of a temporary overloading or maintenance of the server or the server simply wishes to refuse the connection.