CHANGELOG.md

 changelog

0.3.19

• Bump Nginx to version 1.27 from 1.23. and cURL ( 8.9.1-r1 )

0.3.18

• Allow override of client_max_body_size via NGINX_CLIENT_MAX_BODY_SIZE environment variable. Keeping the default to 5MB if not set.

0.3.17

• Allow override of keepalive_timeout via NGINX_KEEPALIVE_TIMEOUT environment variable. Default is 20s (which is different than the Nginx default of 75s) for historical reasons.

0.3.16

• Exclude query params from logging path by default with optional override. Set NGINX_LOGS_INCLUDE_QUERY_PARAMS=true to enable query params in logs.

0.3.15

• added codeowners and bumped to curl 8.5.0-r0

0.3.13/4

• enable master branch publishing to dockerhub

0.3.12

• Added status code access log filtering. Set NGINX_LOGS_INCLUDE_STATUS_CODE_REGEX=^[45] and only see 4xx and 5xx request logs. Default all status codes are logged.

0.3.11

• Moved to slimmer, Alpine base image to shrink the size of the sidecar.

_

• Added boolean publish_to_docker_hub parameter for CirleCI. Defaults to false.
• Added CI/CD test that documents how to use the Sidecar.

0.3.10

• Bumped to nginx:1.23.3 and used script inside of deliveroo/circleci to push multiple tags to Docker Hub.

0.3.9a

• Added :latest tag

0.3.9

• Bumped to nginx:1.23.2

0.3.8

• 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)

0.3.7

• Enabling ngnix status for detailed monitoring.
• New variables are:
• NGINX_STATUS_PORT (default 81) a port to run the status module on
• NGINX_STATUS_ALLOW_FROM (default all) IP, CIDR, all for the nginx config's allow statement (http://nginx.org/en/docs/http/ngx_http_access_module.html)

0.3.6

• Allow specifying the app hostname to proxy to.

0.3.5

• Increase the size of the buffer that stores the response headers to 8K.

0.3.4

• Add $upstream_status to the nginx access log line

0.3.3

• 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.

0.3.2

• Use $request_method and $request_uri instead of the combined $request.

0.3.1

• Log $request_time; request processing time in seconds (millisecond resolution).
• Stop logging unusable information, e.g. private remote IPs, proxy path information.

0.3.0

• Allow client_body_buffer_size to be overridden.

0.2.2

• Add --max-time to the curl health check.

0.2.1

• Wait for the application to pass healthchecks before listening (and accepting) downstream healthchecks.

0.2.0

• Pass original request scheme to consuming application

0.1.0

• Pass original request hostname to consuming application

0.0.1

• Initial version