Forum Discussion

Sara123's avatar
Sara123
Icon for Nimbostratus rankNimbostratus
Jul 10, 2023

How to tackle the downtime for maintenance of nginx

By default the health check time is 5 seconds and 16 s is the time out time for a server added under f5 pool. When we take down one server for maintenance, for 16 s we are seeing alternative request ...
application delivery
  • Nikoolayy1's avatar
    Nikoolayy1
    Jul 11, 2023

    See https://docs.nginx.com/nginx/admin-guide/load-balancer/http-health-check/

     

    http {
    #...    match server_ok {
        status 200-399;
        body   !~ "maintenance mode";
    }
    server {
        #...        location / {
            proxy_pass   http://backend;
            health_check match=server_ok;
        }
    }
}

     

    Here the health check is passed if the status code of the response is in the range 200–399, and its body does not contain the string maintenance mode. 

Paulius's avatar
Paulius
Icon for MVP rankMVP
Jul 10, 2023

Sara123 I am not intimately familiar with nginx but I would imagine they have a similar health monitor option as just a normal F5 device for health monitors which will most likely solve your problem. On an F5 device you have a field in your HTTP health monitor that is a "receive down" string which allows the F5 to mark a pool member as down when a certain string is found in the HTTP response. Clients typically will use a string such as "down" in the HTTP response but this can be adjusted to whatever string you would like. With this health monitor feature being used the pool member should be marked down within 5 seconds because as soon as the F5 sees it the pool member is marked down. I would imagine that nginx will have the same option if you would like to look around that documentation.