Maintenance And Performance Consideration For Cloud Orchestrator

What are maintenance and performance considerations for Cloud Orchestrator.

Issue Description: While using a Cloud Orchestrator and its performance is good for the first few months. But post having millions of logs and hundreds of thousands of queue items, the Instance is quite slower now.
How to improve Cloud Orchestrator performance?

Resolution: Archiving or deleting any data is not to be performed as it will cause data loss in products like "Insights".

  • Archiving or deletion will not affect the Insights functionality in any way as the data is only deleted from Orchestrator and it is already been synced to Insights and will stay available there.

To acknowledge the need for a faster environment, in which operations happen in the blink of an eye, UiPath will be introducing Data Retention and Archiving policies for queue items in Orchestrator, which ensures built-in data off-loading capabilities. Collaboration is required to strengthen operational performance of the Cloud Orchestrator.

This change will impact all Automation Cloud organizations.

The implementation for data retention and archiving policies is broken down into three phases, giving enough time to be informed and to plan the next steps:

Implementation Phases

  • Phase 0 is a pre-deployment phase, all Automation Cloud organizations are informed about the upcoming policies, how the account is impacted, the feature behavior and the rollout mechanism. At the end of phase 0, the feature UI and functionality will be deployed to all cloud environments, but no policy will be activated.
  • Phase 1 is a 6 weeks period of time between the policies deployment and their activation, giving the time to adjust and prepare your queues. Even if the account is prepared before the policies live date, as an Application Information counter will display the remaining days until retention policies apply. A link to the feature documentation, helps to find guidelines about the available policy configuration options. At the end of phase 1, all policies, either the default one or the ones configured, are applied.
  • Phase 2 means policies became active and your cloud account data is offloaded based on their configuration. Phase 2 has no end date. This means that if a new policy is configured, it will apply immediately.

Targeted Resources

The policies are applied to the queue items (and related queue item events and queue item comments) in the associated queues. That means all the queues in the tenant will be mapped to a Data Retention Policy (the default one or the one configured), and this policy applies to the corresponding queue items.

The Policy Mechanism

Here are some highlights to bear in mind when preparing queues for data offload:

  • Configure the policy via the UI while creating a new queue or editing an existing one
  • A new API called Queues-Retention will be available, allowing you to configure the policy programmatically.
  • Choose a maximum of 180 days as the retention duration;
    • Choose a lower duration, but it cannot be higher than 180 days. That is 6 months.
    • The duration refers to queue items that have not been modified for longer than specified.
    • Only queue items in final states will be subject to the policy, this includes those in status Failed, Successful, Abandoned, Retried, Deleted
    • A background job will be regularly checking queue items and their policy and take the necessary actions.
  • Queue items affected by the policy can be either deleted or archived, depending on what is chosen as the outcome:
    • If the policy is set to Delete , all items older than the specified duration will be removed.
  • If the policy is set to Archive , all items older than the specified duration will be archived into an existing storage bucket, that is NOT read-only.
    • To put it briefly, create a, or use an existing, non read-only storage bucket with the main purpose of holding your queue items archive, and then select it.
  • One storage bucket can be used to archive queue items from different queues
    • Use the Orchestrator built-in storage bucket functionality for this purpose, as well as the own external storage bucket.
    • To retrieve the archived information, access the archive files from the designated storage bucket. Publish the format and structure used subsequently in UiPath documentation.
  • The Default Policy keeps the queue items for 120 days and then deletes them. Be careful, as it is not possible to undo deletion.
  • Preserve items’ Unique References, to guarantee the occurrence of validations after the policy applied.
  • Queue item events and queue item comments will be included within the archive.
  • Policy output will NOT affect data synchronization towards UiPath Insights. That is, any board containing the archived or deleted queue items will still be valid and contain their information.


There is a default policy of 120 days+deletion, but it is recommended to not use it. However, set a retention policy. Use the Orchestrator UI or API to configure own retention policy, based on the business requirements.

In case, default policy helps to resolve, do not perform anything specific from this point onward.