Troubleshooting 503 errors

Backend Fetch Failed errors

If the length of cache tags used by Magento exceed Varnish’s default of 8192 bytes, you can see HTTP 503 (Backend Fetch Failed) errors in the browser. The errors might display similar to the following:

Error 503 Backend fetch failed
Backend fetch failed

To resolve this issue, increase the default value of the http_resp_hdr_len parameter in your Varnish configuration file. The http_resp_hdr_len parameter specifies the max header length within the total default response size of 323768 bytes.

1. As a user with root privileges, open your Vanish configuration file in a text editor:

   • CentOS 6: /etc/sysconfig/varnish
   • CentOS 7: /etc/varnish/varnish.params
   • Ubuntu: /etc/default/varnish
2. Search for the http_resp_hdr_len parameter.
3. If the parameter doesn’t exist, add it after thread_pool_max.

4. Set http_resp_hdr_len to a value equal to the product count of your largest category multiplied by 21. (Each product tag is about 21 characters in length.)

   For example, setting the value to 65536 bytes should work if your largest category has 3,000 products.

   For example:

   -p http_resp_hdr_len=65536 \
   

5. Set the http_resp_size to a value that accommodates the increased response header length.

   For example, using the sum of the increased header length and default response size is a good starting point (e.g., 65536 + 32768 = 98304):

   -p http_resp_size=98304 \
   

   A snippet follows:

   # DAEMON_OPTS is used by the init script.
   DAEMON_OPTS="-a ${VARNISH_LISTEN_ADDRESS}:${VARNISH_LISTEN_PORT} \
        -f ${VARNISH_VCL_CONF} \
        -T ${VARNISH_ADMIN_LISTEN_ADDRESS}:${VARNISH_ADMIN_LISTEN_PORT} \
        -p thread_pool_min=${VARNISH_MIN_THREADS} \
        -p thread_pool_max=${VARNISH_MAX_THREADS} \
        -p http_resp_hdr_len=65536 \
        -p http_resp_size=98304 \
    -p workspace_backend=98304 \
        -S ${VARNISH_SECRET_FILE} \
        -s ${VARNISH_STORAGE}"
   

Health check timeouts

If you disable the cache while Varnish is configured as the caching application and while Magento is in developer mode, it might become impossible to log in to the Admin.

This situation could happen because the default health check has a timeout value of 2 seconds. It could take more than 2 seconds for the health check to collect and merge information on every health check request. If this occurs in 6 out of 10 health checks, the Magento server is consindered unhealthy. Varnish serves stale content when the server is unhealthy.

Because Admin is accessed through Varnish, you cannot log in to Admin to enable caching (unless Magento becomes healthy again). However, you can use the following command to enable cache:

bin/magento cache:enable

For more information about using the command line, see Get started with command-line configuration.