MATRIXX Engine Connection Properties

connections:
  siteA:
    host: sitea.example.com
  siteB:
    host: siteb.example.com
    port: 99999

The following gateway.yaml excerpt shows the common configuration default values (where defaults exist) for these properties:

# Settings Related to Matrixx Engine
engine:

  # Is the functionality to connect to the Engine enabled?
  enabled: false

  # Add ttl configuration values
  # To enable set to true
  # Grace period for service call
  # Will use the engine.requestTimeout with engine.ttl.gracePeriod to define aggregate timeout
  ttl:
    enabled: true
    gracePeriod: 2000  

  # Delay (milliseconds) between retries to download MDC definitions
  loadMdcRetryDelayMillis: 5000

  # Optionally disable loading the MDCs from the engine
  loadMdcFromEngine: true

  # MDC Schema and Extension Version. Leave as 0 for the Engine 
  # to decide the level. Specify a value when 'version pinning' during upgrade
  mdc:
    schemaVersion: 0
    extensionVersion: 0

  # Retry requests to the engine if an error occurs (e.g. a connection failure)
  requestRetry:
    # Retry requests
    enabled: true
    # Max number of retries before responding with an error. Set to 0 for unlimited retries.
    # Note: If a connection failure occurs, connections in the pool may not be detected as broken
    # until a request is attempted. So this value should be greater than poolsize.
    maxRetries: 32
    # Interval in milliseconds between retry attempts.
    interval: 100

  # Engine connection details.
  # A "connection" is in fact a pool of TCP connections to a host/port.
  #
  # Depending on the environment, SBA can connect to:
  #   * MDC Gateway
  #   * TRA-LB
  #   * TRA-RT
  #   * Gateway Proxy
  connection:
    # Hostname of one of the above endpoints (e.g. TRA-RT)
    # Note: This will be ignored if hosts are specified in the connection.connections section.
    host: localhost
    # Port to connect to
    port: 14060
    # Authenticated port (of Gateway Proxy). If connecting over the Authenticated port of
    # Gateway Proxy, set Port to the same value as this.
    authPort: 14080
    # Timeout when creating a connection
    connectTimeout: 10000
    # Timeout duration each request
    requestTimeout: 5000
    # Send a ping at this interval (in seconds). Set to zero to disable.
    pingInterval: 10
    # Max number of successive failed Ping messages before reconnecting
    maxPingFailures: 2
    # The value of the TrafficRouteData field of the MtxRequestSysQueryDomain message used 
    # to ping a connection. By default this is set to RTID0 which will route to any
    # available subdomain. In multi-domain deployments this may need to be changed to
    # specify a particular subdomain (in 523x) as RTIDx, or domain id (524x onward) as
    # DMIDnn. In 524x, the Helm Chart will configure this automatically if a Domain is 
    # specified.
    pingRouteId: RTID0
    # If connecting via the Authenticated port of Gateway Proxy, create login credentials
    # in Gateway Proxy for SBA Gateway and specify those credentials here.
    credentials:
      username: sba_username
      password: sba_password
      apiKey: sba_apiKey

  # Connect to multiple hosts. For each (pool of) connections, a host must be specified.
  # The default values for each connection pool will be taken from the connection configuration
  # section above. Each value can be overridden per connection pool.
  # e.g:
  # connections:
  #   siteA:
  #     host: sitea.example.com
  #   siteB:
  #     host: siteb.example.com
  #     port: 99999
  #
  # NOTE: For Kubernetes deployments, a single connection will be routed over a Service,
  #       which will provide load balancing and failover to multiple TRA-RT instances.
  #       This configuration option is more likely required for hybrid deployments
  #       where the TRA-RT is not inside the Kubernetes cluster.
  connections: []

  # Failover mode when multiple connections are defined.
  #   ACTIVE_STANDBY (default): Use the first connection in the list of connections unless it fails. On failure,
  #                             use the next in the list and so-on.
  #   ACTIVE_ACTIVE: Balance
  # Note: If pingInterval is defined, the ping message will be used to determine if
  #       a connection pool is available again. If pingInterval is set to zero, 
  #       incoming requests will try to use a pool even if it has previously failed.
  failoverMode: "ACTIVE_STANDBY"