Release date: 20th January 2020
With this release we are excited to introduce you to the newest member of the UiPath family, Insights. A powerful embedded analytics solution integrated within Orchestrator at the tenant level, Insights helps you measure, optimize and scale your RPA operations and its impact to your business.
Gain operational awareness with a library of out of the box dashboards, customizable with any combination of widgets and/or filters. Dashboards can be shared with any or all users of the Orchestrator tenant, enabling company-wide visibility while continuing to work from the same dataset and without impacting your original dashboards.
Using the Pulse feature in Insights, you can stay up-to-date with the current status of your RPA deployment’s most important metrics and receive alerts of any errors or anomalies.
So as to better label Robots that do not rely on an Orchestrator license, all license source fields in Orchestrator that depict this concept have been renamed using the Stand-alone notion.
We added a new
web.config parameter which allows you to read logs from targets that aren't necessarily used to write logs into, namely
Logs.RobotLogs.ReadTarget. We did this for those of you who use different methods to push logs other than through the
In this release, a new
web.config parameter has been added which enables you to convert local users that log in with AD credentials into directory users:
WindowsAuth.ConvertUsersAtLogin. Details here. Since this parameter doesn’t make changes retroactively, we also put together a script to convert all local users that have ever logged in with AD credentials into directory users at once. See it here.
Changes made to AD group membership are now reflected in the Robot tray after each successful Robot connection or once every hour. This value can be changed using the
WindowsAuth.GroupMembershipCacheExpireHours parameter in
Active Directory users and groups can now be assigned to folders without first having been added globally as Orchestrator users. See how here.
To make searching for an AD user or group quicker and easier, we made a couple of tweaks to the original recipe to boost the performance and improve the overall experience. The highlights you need to remember:
- We added a User Type drop-down which enables you to refine the query by selecting the type of entity you're searching for (user, group, or both)
- We added three
web.configparameters which enable you to make finer adjustments based on your needs: one to control the maximum number of results returned for a query, one to define the minimum query size which triggers the search, and one which limits the search duration. Details about all three here.
Making an authenticated API request using an Active Directory identity that belongs to a group which has been added in Orchestrator beforehand auto-provisions the user in your instance. Be aware of the fact that you cannot choose the tenant the user is auto-provisioned on, so if he belongs to multiple tenants, auto-provisioning takes place on one of them only.
We made some tweaks to the logic behind the job ordering mechanism in Orchestrator so as to optimize the execution in some corner cases. More precisely, jobs are now added to the workload and then executed based on their creation time. This is valid regardless of the execution target. Since we’re here, you should also know that we improved the performance when using dynamic allocation.
The User Permissions window has been enhanced to make it easier to assess what permissions your user has been granted on a per folder basis, and whether they have been explicitly assigned or inherited from the tenant.
The Include Subfolders option is now also available for the SLA page as well, such that you can select if the contents of any existing subfolders are included or not when monitoring your queues.
Managing your organization and deployment using modern folders is made easier as you can now automatically create roles for each type of user. See the corresponding permissions for each available role here.
Having in mind to offer you a more intuitive user experience, we've revamped the Non-Working Days tab in the Settings page. As a result, now the calendars are alphabetically ordered, the selected calendar is clearly marked, and the buttons have tooltips.
All alert messages now contain the full folder path to help you easily identify where the error occurred.
- If using a self-signed certificate, Insights is inaccessible from Firefox unless you explicitly select to continue to the site. See here for details.
- When performing a new installation of Insights and Orchestrator with Insights feature enabled against a SQL database where Insights was previously installed and configured, the Orchestrator will display an error when the Insights page is accessed. This is corrected by running the
UiPath.InsightsTenantAdminTool.exeas detailed here.
- When performing a modify/repair installation to your existing Insights deployment, entering the wrong credentials for the Insights admin user will lock the account and render it inaccessible for 10 minutes.
- When configuring a Pivot widget, if a large set of row values is set as columns this can result in both Orchestrator and Insights freezing, which can only be corrected by restarting the browser.
- Changes in the Orchestrator localization settings are not propagated to the Insights Portal until you logout and back in to Orchestrator.
- Some modifications to the processes and/or queues whose data is displayed on any dashboard may result in that dashboard’s temporary unavailability. See here for details.
- Creating and editing of dashboards on a low-resolution display does not display the full range of controls unless you zoom out.
- Process workflows where all custom log fields are removed during execution are not ingested into the Insights database. See here to read about considerations when designing workflows for Insights.
- Insights users with the same email address used across multiple Orchestrator tenants have access to data from across all of those tenants.
- Upon upgrading your existing Orchestrator instance to v2019.10, certain scenarios occurred in which organization units were not properly displayed as the equivalent classic folders, thus rendering Orchestrator unusable. Specifically, if you had multiple OU's in a previous Orchestrator version, and deleted at least one which had a user assigned to it, the folder correspondents of the remaining units were not visible to the respective user after migrating.
- Users assigned to non-sequential folders in a hierarchy (e.g. Root and Subfolder 6) were unable to access the intermediate folders.
- Fixed an issue where the folder structure (i.e. root and subfolders) was not taken into consideration for the alphabetical ordering of the Folders list. For example, with root folders Accounting and Accounting & Finance, where Accounting has subfolders Payroll and Reports, Accounting & Finance was displayed directly under Accounting rather than it's subfolders.
- The Robot column was not populated for transaction items processed by Robots in modern folders. As of now, the column displays the user associated to Robot which processed the items.
- Users with Create permissions on Subfolders were unable to create any subfolders unless they were also granted Create permission on Folders.
- Having a large number of Robots (in the order of thousands) in a single environment caused a significant performance degradation when dealing with processes. We improved the logic around process handling to prevent that from happening.
- We identified and fixed an issue related to the permissions required to configure automatic robot creation, and to receive alerts on robot status changes.
- The roles and folders inherited by a directory user at login time were not audited.
- Alerts concerning automatically provisioned Robots in classic folders did not contain the Robot name.
- Exceptions were thrown in Event Viewer when you tried to open the Alert Details window for alerts on transactions not contained in the currently active folder. As a result, the Show Details button has been disabled for transaction alerts in this scenario. Switch to the containing folder to see the alerts that have been generated.
- On the Logs page, filtering by User did not allow for local users and incorrectly enforced case-sensitive searching.
- We identified an issue with the disabling time of a queue trigger when also having a time trigger associated to that same queue. This only occurred if the latter had been configured to get disabled 15 minutes earlier than the first one. Say I defined a time trigger for a queue to get disabled at 10:00 o'clock, and a queue trigger for the same queue which should get disabled at 10:15. What happened is that the queue trigger got disabled at 10:00 instead of 10:15.
- The values of input and output parameters defined when creating a queue trigger were reset to their defaults when editing the trigger.
- Output and Analytics schema validation interfered with the retry mechanism of failed transactions.
- For Queues with schemas enforced for Output and/or Analytics data, queue items that contained no Output and/or Analytics parameters were nonetheless processed successfully.
- Schemas using the
rootdefinition rather than
draft-07were detected as invalid.
- When editing or cloning queue items, the new item postpone or due date was entered in UTC rather than the time zone configured for the tenant.
job.startedwebhook event was not sent to an external system for jobs started from the Robot tray.
- Various error messages where displayed when you searched for a directory group or user on the Add Directory User or Group window using special characters such as
- If the Orchestrator to Redis connection was down, an error loop would occur where Orchestrator was inaccessible.
- The Connection String was not properly displayed in the Settings page, as it contained two slashes (
- Fixed an issue that prevented you from downloading the Orchestrator offline activation request file if you were using Internet Explorer.
- Occasionally, a blank page was displayed when trying to authenticate via Windows Authentication on Google Chrome.
- It is now possible to remove a license at tenant level if it was initially uploaded at host level.
- The Interval filter for the Jobs page did not work properly if you used a time zone different than UTC.
- It was not possible to update a host license a second time if you upgraded your Orchestrator instance to 2019.10.x from an unsupported version such as 2018.3.x.
- Auto-provisioned users could not be assigned to tasks in a folder despite belonging to an AD group which had access to it, unless they had been explicitly granted permissions to the respective folder. See here details about assigning a task to a user.
ProvisionTypeparameter was required when making POST requests to the
/odata/Robotsendpoint. We fixed it so that’s no longer the case.
- You could not use Orchestrator if Active Directory was unreachable, even if it was deployed on a machine that was not an Active Directory member.
- When deactivating the license of your offline Orchestrator instance, Orchestrator would remain licensed until an internet connection was enabled.
- If a trigger had a calendar attached to it and the tenant's timezone was changed after calendar creation, then you couldn't edit the trigger unless you first changed its timezone to match that of the tenant.
- The calendars are displayed correctly now in Firefox.
- While using Internet Explorer, you couldn't add new calendars within tenants that had no previously defined ones.
- There were two Save and Delete buttons for each calendar. To avoid confusion, we've reworked the UI, leaving only one of each type. Now the Save button enables you to update both the name and the dates of any selected calendar, while the Delete button removes the calendar entirely.
- Switching between calendars or adding new ones didn't default the view to the current month when the previously selected calendar had dates in another month.
- You couldn't save a calendar if you've previously entered more than the maximum limit of characters allowed for its name, even after shortening the name.
- For tenants with a timezone different than the default (UTC) Coordinated Universal Time, the dates uploaded to a calendar via
.csvfiles were incorrectly displayed with a difference of minus one day.
- The dates uploaded into calendars through
.csvfiles were not saved.
Updated about a month ago