Multiplicators of load balancing criteria. Load balancer decides which cluster node executes graph. It means, that any node may process request for execution, but graph may be executed on the same or on different node according to current load of the nodes and according to these multiplicators.
The higher number, the higher relevance for decision. All multiplicators must be greater then 0.
Each node of the cluster may have different load balancing properties. Any node may process incoming requests for transformation execution and each may apply criteria for loadbalancing in a different way according to its own configuration.
These properties aren't vital for cluster configuration - default values are sufficient
Table 30.1. Load balancing properties
| property | type | default | description |
|---|---|---|---|
| cluster.lb.balance.running_graphs | float | 3 | Specify importance of running graphs for load balancing. |
| cluster.lb.balance.memused | float | 0.5 | Specify importance of used memory for load balancing. |
| cluster.lb.balance.cpus | float | 1.5 | Specify importance of number of CPUs for load balancing. |
| cluster.lb.balance.request_bonus | float | 2 | Specify importance of the fact, that the node is the same which processes request for execution. The same node, which decides where to execute graph. If you specify this multiplicator great enough, it will cause, that graph will be always executed on the same node, which processes request for execution. |
| cluster.lb.balance.node_bonus | float | 1 | Overall ratio bonus for configured node. Values greater then "1" increase probability the node will be chosen by the loadbalancer. Value "1" means no bonus or penalty. "0" means that the node will be never chosen by the loadbalancer, however it still may execute graphs, e.g. when there is no other node in cluster, or when the graph is designed to run on the node. |