- Enabling the support of caching and websockets
- Some variables were renamed, see
entrypoint.sh
- New variables are:
NGINX_CLIENT_BODY_BUFFER_SIZE
- it sets client_body_buffer_size (default: 8k).NGINX_CLIENT_MAX_BODY_SIZE
- it sets client_max_body_size (default: 5M).APP_SCHEME
- it sets application scheme (default: http).NGINX_PROXY_BUFFER_SIZE
- it sets proxy_buffer_size (default: 8k).
- Enabling the support of custom timeout for proxy requests
- New variable is:
PROXY_TIMEOUT
- it sets proxy_connect_timeout, proxy_send_timeout, proxy_read_timeout values. (default: 60s)
- Enabling ngnix status for detailed monitoring.
- New variables are:
NGINX_STATUS_PORT
(default81
) a port to run the status module onNGINX_STATUS_ALLOW_FROM
(defaultall
) IP, CIDR,all
for the nginx config'sallow
statement (http://nginx.org/en/docs/http/ngx_http_access_module.html)
- Allow specifying the app hostname to proxy to.
- Increase the size of the buffer that stores the response headers to 8K.
- Add
$upstream_status
to the nginx access log line
- Add
--fail
to the curl health check. This causes curl to return non-zero exit codes even if the http request completes but the response code represents an error. - Add
--verbose
to the curl health check. This helps us debug any application responses if the server does start, but with errors.
- Use
$request_method
and$request_uri
instead of the combined$request
.
- Log
$request_time
; request processing time in seconds (millisecond resolution). - Stop logging unusable information, e.g. private remote IPs, proxy path information.
- Allow
client_body_buffer_size
to be overridden.
- Add
--max-time
to the curl health check.
- Wait for the application to pass healthchecks before listening (and accepting) downstream healthchecks.
- Pass original request scheme to consuming application
- Pass original request hostname to consuming application
- Initial version