---
title: Using load balancing
description: Use load balancing to run different branches in a flow on different executions.
component: davinci
page_id: davinci:flows:davinci_using_load_balancing
canonical_url: http://docs.pingidentity.com/davinci/flows/davinci_using_load_balancing.html
revdate: July 29, 2024
section_ids:
  steps: Steps
  choose-from: Choose from:
  choose-from-2: Choose from:
  choose-from-3: Choose from:
---

# Using load balancing

Use load balancing to run different branches in a flow on different executions.

## Steps

1. Click the **Flows** tab.

2. Locate and open the flow.

3. Click the **Action Decision** node.

4. Select **Load Balance** to enable load balancing.

5. In the **Load Balance Context** list, select a context for load balancing.

   The context determines how widely the algorithm is applied. For example, a Round Robin algorithm with a local context will always begin with the first option and only use the second option if the flow returns to the same node again. However, a Round Robin algorithm with the Global context will consider which option was used in the previous flow execution.

   ### Choose from:

   * **Global**: The algorithm is applied using data from all flow executions.

   * **Local**: The algorithm is applied using only data from the current flow execution.

6. In the **Load Balance Algorithm** list, select an algorithm to use for load balancing.

   ### Choose from:

   * **Random**: Uses a randomly selected exit path.

   * **Random Weighted**: Uses a randomly selected exit path, with the probability of each path determined by its weight.

   * **Round Robin**: Cycles through the available exit paths, with each execution using the option after that of the previous execution.

7. For each path out of the node, select a logical operator to determine when it should be considered by the load balance algorithm.

   ### Choose from:

   * **All Triggers True**: All of the nodes leading into the action decision node completed successfully.

   * **All Triggers False**: All of the nodes leading into the action decision node failed.

   * **Any Triggers True**: One or more of the nodes leading into the action decision node completed successfully.

   * **Any Triggers False**: One or more of the nodes leading into the action decision node failed.

   * **Any Trigger Completes**: One or more of the nodes leading into the action decision node completed either successfully or unsuccessfully.

   * **All Triggers Complete**: All of the nodes leading into the action decision node completed either successfully or unsuccessfully.

8. If you selected the **Random Weighted** algorithm, select a weight for each path out of the node.

9. Click **Save**.