Queue Types (Routing Algorithm)
    • Dark
      Light
    • PDF

    Queue Types (Routing Algorithm)

    • Dark
      Light
    • PDF

    Article summary

    Rocket.Chat offers diverse types of queues with unique routing algorithms to efficiently manage and distribute incoming Omnichannel conversations in your workspace. They define the routing methods and direct conversations to the appropriate agents.

    You can manage omnichannel queues in the Routing settings of your workspace.

    Type of Omnichannel queues

    Here are various types of queues you can set for routing in your Omnichannel workspace:

    Auto selection

    This round-robin algorithm is designed to automatically distribute chats evenly among agents on the workspace. It is especially suited for operations focused on equal opportunity distribution, as it prioritizes agents based on the total number of chats they have handled since their initial login. If an agent goes offline, they won't get new chats until they're back online. Their chat count stays the same during their offline period. Once they return, the algorithm prioritizes assigning new chats to them until they've caught up with their teammates in terms of the total number of chats served.

    Manual selection

    When manual selection is enabled, each agent has a Queued Chats section where new incoming chats are displayed. Agents can click on a chat to preview it, view the messages sent, and decide whether or not to take the chat. Once an agent takes a chat, it is removed from the Queued Chats section for all agents. With manual selection active, agents can see all conversations in the queues associated with their department(s). This visibility allows them to preview chats and gather the necessary context for providing efficient and satisfactory service. Department-specific chats are only displayed in the Queued Chats section for agents within that department.

    External service

    To integrate your custom agent routing rule into Livechat, use an External Service. Once it's configured as the Livechat routing method, update the External Queue Service URL and Secret Token in the Routing settings.

    Rocket.Chat sends a GET request to the External Queue Service URL and the Secret Token is sent as a header X-RocketChat-Secret-Token for you to validate if the request came from the Rocket.Chat. If Rocket.Chat receives a response status other than 200, it will retry up to 10 times until a valid response is received.

    After submitting the GET request, Livechat expects a JSON response in the following format:

    {
         "_id": "CbbQkRAifP6HtDLSr",
         "username": "valid.username"
    }

    Once the valid response is received in the above format, Livechat verifies that the provided username represents a valid Livechat agent, proceeding with the standard process flow.

    Load balancing

    The load-balancing algorithm distributes chat workloads based on the number of ongoing chats per agent. It aims to balance the workload by considering the current chat volume each agent is handling. Unlike auto selection, which distributes chats based on the total number of chats served, load balancing focuses on the current workload to maintain an even distribution of active chats among agents.

    Load rotation

    The load rotation algorithm assigns chats based on when an available agent last received a chat, focusing on the time of the last assignment rather than the current workload. It ensures that chats are distributed in a rotating sequence among all available agents, promoting a fair and balanced distribution for everyone.

    Workspace administrators can leverage the various types of queues outlined to achieve dynamic load balancing, strategic auto-selection, and effective chat handling, tailoring the Omnichannel experience to their specific organizational needs to enhance productivity.

    The next topics describe the Omnichannel features and settings in detail.


    Was this article helpful?

    What's Next
    ESC

    Eddy AI, facilitating knowledge discovery through conversational intelligence