Services

An ingress configuration file may define zero or more services. Each service groups a set of routes with one or more upstreams and an optional load balancing strategy.

services = [
  {
    # Load balancing strategy for this service.
    load_balancing_strategy = "round_robin"

    # Optional: health checking for upstreams.
    health_check = {
      enable                     = false
      failure_threshold          = 3
      unhealthy_cooldown_seconds = 10
    }

    # Optional: circuit breaker for upstream protection.
    circuit_breaker = { ... }

    # Routes and upstreams are defined within the service.
    routes    = [ ... ]
    upstreams = [ ... ]
  }
]

Fields

Field	Type	Default	Description
load_balancing_strategy	string	"failover"	The load balancing strategy. See supported strategies below.
health_check.enable	boolean	true	Enable health checking for upstreams.
health_check.failure_threshold	integer	5	Number of consecutive failures before marking an upstream as unhealthy.
health_check.unhealthy_cooldown_seconds	integer	10	Seconds to wait before re-checking an unhealthy upstream.
circuit_breaker	object	(optional)	Circuit breaker configuration. See Circuit Breaker.
routes	list(object)	(required)	Route matching rules. See Routes.
upstreams	list(object)	(required)	Backend upstream servers. See Upstreams.

Load Balancing Strategies

• failover: Always picks the first healthy upstream in the list.
• round_robin: Distributes requests evenly across upstreams. Respects the weight field on each upstream.
• request_pressure: Picks the upstream with the lowest recent request pressure (heuristic-based, not transport-level).
• random: Picks a random healthy upstream.
• sticky_hash: Consistent hashing based on request characteristics.