Behaviour:

• max-attempts
  • Httpd
    • Amount of failover requests to perform if the first request fails
  • Undertow
    • Amount of request to perform(first request + failover requests)

• max-retries
  • Httpd
    • Not present
  • Undertow
    • Amount of failover requests to perform if the first request fails
• max-retries vs max-attempts (undertow balancer)
  • Attribute containing bigger value is used, if values are equal then "max-retries" is used
  • max-retries is set globally for all balancers
  • max-attempts is set for individual balancer

Documentation:
Undertow subsystem

            "max-retries" => {
                "description" => "The number of times to attempt to retry a request if it fails. Note that if a request is not considered idempotent then it will only be retried if the proxy can be sure it was not sent to the backend server).",
            },

Mod_cluster subsystem

            "max-attempts" => {
                "description" => "Number of times the reverse proxy will attempt to send a given request to a worker before giving up.",
            },

Upstream documentation

maxAttempts - Maximum number of failover attempts before giving up. The minimum value is 0, i.e. no failover. The default value is 1, i.e. do a one failover attempt.

Issue:

• No information about behavior when max-retries == max-attempts .
• Cli documentation in mod_cluster subsystem describes only scenario when undertow balancer is used