About Queues and Transactions

A queue is a container that enables you to hold an unlimited number of items. Queue items can store multiple types of data, such as invoice information or customer details. This information can be processed in other systems - SAP or Salesforce, for instance.

The data stored in, and output from, Queue items is free form by default. For situations where a specific schema is needed, such as integrations with other applications, processing of machine generated forms, or for analytics, you can upload custom JSON schemas to ensure that all Queue item data is in the proper format.

In Orchestrator, newly created queues are empty by default. To populate them with items you can either use the upload functionality in Orchestrator, or Studio activities. The latter also enable you to change item statuses and process them. As soon as queue items are processed, they become transactions.

Queues Overview

Queues enable you to create large automation projects underlined by complex logic. For example, you can create a process that collects all invoice information and creates a queue item for each piece of data to store it. Subsequently, you can create another process that gathers the information from Orchestrator, and uses it to perform additional tasks, such as paying the invoices in a different application, postponing their payment according to their due date or value, sending emails to the accounting team every time a bill is paid, etc.

The Queues page enables you to create new queues, to view previously created queues, and to access charts displaying the transaction status progress over time, along with various other details, such as average execution time and the total number of successful transactions.

The data available in the queues grid is updated at regular intervals, which means that it is not always displayed in real time, and that slight delays are possible. Additionally, it is not affected by retention policies, so archiving database items does not change the information available in the grid.

Item statuses are controlled by RPA developers when they create the automation projects, while revision statuses are controlled in Orchestrator and enable you to perform version control, but only of queue items that have been abandoned or have failed with an application or business exception.

Failed or abandoned items can also be assigned to a reviewer, which can be changed or cleared at any point, if needed. Each of these changes are tracked in the History tab of the Audit Details window. The reviewer is in charge of assessing the current status of the transactions he is assigned to, and changing the review status. The status of queue items up for revision can be changed in the Review Requests page.

Important: To ensure that queue item processing is performed successfully, you need to have a unique reference set at the queue level when dequeuing by reference. Failing to keep the reference unique will lead to concurrency access issues, such as failures due to no transaction data being found (

No
                     Transaction Data

error message).

If the reference cannot be unique, it is advisable to dequeue without reference.

Schema Definitions

When creating or editing a queue, you can upload a custom JSON schema for the Specific Data,Output Data, and/or Analytics Data. With the schema(s) in place all transactions are validated against the provided format, and if the resulting data does not conform that item fails with a Business Exception.

Important:

The schema is not applied retroactively to existing transactions, only to those executed after you have uploaded the schema(s).
Your schema(s) must not contain an array.
For validation purposes, DateTime is accepted as string type.
Use of and validation of an Analytics Data schema requires Robots and Activities of version 19.10 or greater.
If the uploaded schema(s) do not contain a valid schema definition URI, then the draft-07 definition, as in the below example, is used as the fallback.

Note:

The Specific Data size of queue items is limited to 1 MB to achieve better control in terms of Orchestrator performance. Anything beyond this limit cannot be added to a queue, and it returns the `403 - Payload Too Large` error code. If you need to upload larger items, store the large data in an external storage and only reference the link in the item.

A sample schema:

{
  "definitions": {},
  "$schema": "http://json-schema.org/draft-07/schema#",
  "$id": "http://example.com/root.json",
  "type": "object",
  "title": "The Root Schema",
  "additionalProperties": { "type": "string" },
  "required": [
    "stringTest",
    "intTest",
    "boolTest"
  ],
  "properties": {
    "stringTest": {
      "$id": "#/properties/stringTest",
      "type": "string",
      "title": "The Stringtest Schema",
      "default": "",
      "examples": [
        "stringTest"
      ],
      "pattern": "^(.*)$"
    },
    "intTest": {
      "$id": "#/properties/intTest",
      "type": "integer",
      "title": "The Inttest Schema",
      "default": 0,
      "examples": [
        30
      ]
    },
    "boolTest": {
      "$id": "#/properties/boolTest",
      "type": "boolean",
      "title": "The Booltest Schema",
      "default": false,
      "examples": [
        false
      ]
    }
  }
}{
  "definitions": {},
  "$schema": "http://json-schema.org/draft-07/schema#",
  "$id": "http://example.com/root.json",
  "type": "object",
  "title": "The Root Schema",
  "additionalProperties": { "type": "string" },
  "required": [
    "stringTest",
    "intTest",
    "boolTest"
  ],
  "properties": {
    "stringTest": {
      "$id": "#/properties/stringTest",
      "type": "string",
      "title": "The Stringtest Schema",
      "default": "",
      "examples": [
        "stringTest"
      ],
      "pattern": "^(.*)$"
    },
    "intTest": {
      "$id": "#/properties/intTest",
      "type": "integer",
      "title": "The Inttest Schema",
      "default": 0,
      "examples": [
        30
      ]
    },
    "boolTest": {
      "$id": "#/properties/boolTest",
      "type": "boolean",
      "title": "The Booltest Schema",
      "default": false,
      "examples": [
        false
      ]
    }
  }
}

Transactions Overview

The Transactions page displays the transactions from a given queue. It also shows their statuses, the dates when they should be processed, the Robot that processed them, and the type of exception thrown or assigned reference, if any.

You can search for a specific transaction or a group of them, according to a custom reference, which is added through the Reference property of the Add Queue Item and Add Transaction Item activities. The reference can be used to link your transactions to other applications used within an automation project. Additionally, this feature enables you to search for certain transactions in Orchestrator, according to the provided custom reference.

Transaction references can also be enforced to be unique, at queue level. This feature is enabled when creating the queue and applies to all transactions except deleted or retried ones. This makes identifying a specific item a breeze and eases the review process.

If a duplicate reference is encountered while adding items to a queue, the job fails with a Faulted status and displays the Execution error: UiPath.Core.Activities.OrchestratorHttpException: Error creating Transaction. Duplicate Reference. error message in the Job Details window.

Processing Order

Within any given queue the transactions are processed in a hierarchical manner, according to this order:

Items that have a Deadline, as follows:
- in order of Priority; and
- according to the set Deadline for items with the same Priority.
Items with no Deadline, in order of Priority, and
- according to the rule First In, First Out for items with the same Priority.

When setting a Deadline or a Postpone date, we recommend populating the respective fields with relative dates. For example, DateTime.Now.AddHours(2), DateTime.Now.AddDays(10) and DateTime.Now.Add(New System.TimeSpan(5, 0, 0, 0)). Additionally, you can use the US notation to add an exact time, such as 10/10/2019 07:40:00. Automatic correction of this date is available, for example, if you write 12 10 2019 9:0, it is automatically transformed to 12/10/2019 09:00:00.

Deadline is useful for ordering tasks with similar priorities, while Postpone ensures that a task is not initiated before the specified time. However, the two parameters are not designed to be used together.

The dates added in Studio for the Deadline and Postpone fields are displayed in Orchestrator, in the Transactions page, under the Deadline and Postpone columns.

Exporting Transactions

You can export all the transactions and information related to a given queue to a .csv file, by clicking the Export button, in the Transactions page. All page filtering options apply to the generated file, too.

To ensure the best performance, please note that the exported entries are not in reverse chronological order.

Queue SLA Predictions

This tool helps you set an SLA (item Deadline) for newly added items in a queue. This helps you assess if they can be processed in a timely manner, and what resources you need to allocate such that their SLA is not breached. Whenever the SLA is in danger of not being met, you are properly notified such that you can make adjustments accordingly.

The SLA only applies to those items that don't have a deadline set, meaning that a newly added item with no deadline defined beforehand has it automatically filled in according to the value set as the SLA. Specifically, each item's deadline is represented by the value set for the queue SLA from the moment the item was added into the queue. For instance, if I set the SLA to 2 hours, and I add 3 items into the queue at 4, 5, and 6 PM, then my items have the deadlines 6, 7, 8 PM, respectively.

The SLA breach alert is triggered every 30 minutes, starting at 7 minutes past the hour.

Items which have a Deadline (either set in Studio or in a .csv file used for upload) are not affected by the SLA setting.

Important:

The Priority of the items added in a queue after enabling SLA predictions is automatically set to High, regardless of how it was set in Studio or in the .csv file used for upload.
You cannot delete a process associated to a queue with enabled SLA predictions.
If at least one queue item exceeds its deadline, Over Capacity is displayed in the Necessary Robots (SLA) column and predictions are no longer calculated.
Predictions are made for queue items with deadlines in the next 24 hours, and do not take into account the items' defer dates.

Queue triggers and SLA predictions are interdependent in terms of queue-process association. So whenever configuring one, the other is prefilled such as to have parity between the configurations. Say I define a queue trigger for queue Y to use process X. SLA predictions for queue Y can only be made using process X, therefore X is prefilled and read-only when enabling queue SLA for Y.

Risk SLA

You can also define a Risk SLA for your items, which works like a buffer zone before the SLA. Explicitly, the risk deadlines of your items are calculated based on the Risk SLA from the moment the queue item was added in the queue. Say I set the Risk SLA to 2 hours, and I add 3 items into the queue at 4:30, 5:15, and 6:45 PM, then my items have the risk deadlines 6:30, 7:15, 8:45 PM, respectively.

After the Risk SLA has passed and the queue item is not processed, the item becomes at risk of not meeting its deadline. The user is properly notified, such that he can make adjustments accordingly.