HAProxy Maintenance Page

haproxy_maintenance.conf

#
# Proof of concept for a HAProxy maintenance mode
#
#
# Control the maintenance page during runtime using the stats socket:
#
# To put the whole site in maintenance mode (for all IPs):
# > add acl #0 0.0.0.0/0
#
# To exclude your own ip, so you are able to test things out:
# > add acl #1 8.8.8.8
#
# To exit maintenance mode again simply remove the mask from #0 again:
# > del acl #0 0.0.0.0/0
#

frontend web
    bind 0.0.0.0:80
    mode  http
    default_backend web_backend
    # use this in case of real errors
    errorfile  503 /etc/haproxy/errors/maintenance.http
    # maintenance mode - disabled by default
    acl web_maintenance src --
    acl web_maintenence_ex src --
    use_backend web_maintenance if web_maintenance !web_maintenence_ex
  
  
backend web
    mode  http
    balance  roundrobin
    server web01 127.0.0.1:8080 inter 500 rise 3 fall 1 weight 100 slowstart 30s
  

backend web_maintenence
    mode http
    errorfile  503 /etc/haproxy/errors/maintenance.http

Just saying, there is a typo here, the backend is called "web_maintenEnce", but the use_backend directive references "web_maintenAnce".

Hi How should I use this?

The idea behind is completing the acl in runtime, so web_maintenance is mapped to ACL #0 and web_maintenence_ex to #1. Given so, once running, by executing the lines shown, you can control the behavior of the acl, thus changing the way haproxy process requests.