Refer to the listed KB article for details.

1. Open a Zurich instance with multiple nodes.
2. Upgrade the instance from Zurich to Australia RP2.
3. Run a quick Discovery or discover a few IPs.

Notice that the Shazzam probe creation or Shazzam sensors processing has an error. The Shazzam probe isn't created with a name (likely skipping some business rules), and Shazzam sensor processing doesn't create the further probes in the chain. This results in discovery not running effectively.

1. Navigate to any relevant workspace (for example, SOW or CSM) that contains an Activity Stream.
2. Open a record (for example, Incident) that contains a comment or work note posted by an AI Specialist.

Expected behavior: The 'reviewed by' subtitle isn't displayed on AI Specialist posts.

Actual behavior: The 'reviewed by' subtitle is displayed on all AI Specialist posts (for example, 'Work note reviewed by Abel Tuter').

1. Navigate to any relevant workspace that contains an activity stream.
2. Open a record that contains a comment or work note that an AI Specialist posted.

Expected behavior: There should be a gradient border ring on AI Specialist avatars.

Actual behavior: The gradient border ring is missing from AI Specialist avatars.

1. Open any instance with RTE turned on (glide.ui.journal.use_html=true).
2. Navigate to SOW.
3. Open any incident.
4. Open DevTools and throttle CPU to 6x slowdown.
5. Navigate to Work Notes.
6. Begin typing very rapidly.

Expected behavior: All keystrokes should be recorded in RTE.

Actual behavior: The cursor backups to a previous spot and users lose text.

1. Log in as admin.
2. To configure the instance for participation in an AI specialist workflow, set the system property sn_aia.enable_ai_workers to true.
3. Create a user account with the display name = AI Specialist.
4. Impersonate AI Specialist.
5. Add a new incident record.
6. Copy the sys_id for the record (INCIDENT_SYS_ID).
7. Add two work notes for the incident record.
8. End the impersonation (go back to admin).
9. Look up the sys_id of the first work note (WORK_NOTE_SYS_ID) in the sys_journal_field table.
10. Copy that sys_id.
11. Run a background script to flag the work note for review.
12. Impersonate an ITIL user.
13. Open the incident record in UI16.

Expected behavior: Page level notification is displayed: 'AI Specialist has updated fields and work notes. Please review the work notes and fields for accuracy.' The notification also contains a link titled Review work notes. When this link is selected, configured work note (WORK_NOTE_SYS_ID) is highlighted in the activity stream using citation highlight snake animation.

Actual behavior: No notification is displayed.

1. Set up an off-glide conversation server with a lab instance.
2. Start a chat.
3. Request a live agent.
4. Check the node logs for the instance.

Observe that several error logs are printed, including 'Failed to get client for base url'. The errors are related to unneeded messages being attempted to an incorrect URL.

1. Provision an instance with glide.awa and nowassist-ai-agents plugins on a Z or A instance.
2. Enable 'Include AI Agents' for a assignment rule which has both AI agent and manual agents.
3. Start a incident.

Observe that incident is not assigned to an AI agent as preference even though it is available.

1. Provision a Zurich instance with sn_aia and interaction.awa plugins installed.
2. Navigate to the awa_assignment_rule table.
3. Open or create a entry.

Observe that the Include AI agents related fields are available in the 'Auto-assign handling' section.

1. Provision an instance with Agent Chat installed.
2. Log in as both an agent and requester.
3. Make the agent available.
4. Navigate to Service Portal.
5. Request a live agent.

Observe that the conversation ends without a work item sent to the agent.

Connect to LA.

Observe that the VA Message says 'Live Agent Engaged' before the agent accepts the workItem.

Scenario 1:

Scenario 2:

1. Navigate to sn_aia_property table and set the 'Enable value overrides' property to true.
2. Create a basic agent with two tools.
   1. Create a skill tool.
   2. Create a script tool (a tool to use the skill tools output as input).
3. Set value overrides on the script tool.
4. Create an input, use the value override field to set it to the skills response. The script can be very minimal.
5. Save and run agent in playground.

Expected behavior: The script return the expected skill output that was passed to it.

Actual behavior: The script tool returns undefined, and the agent fails.

1. Navigate to SOW.
2. Impersonate a service desk manager. For example, a user with the role sn_sow_itsm_common.sn_service_desk_manager.
3. Make a copy of the AI Service Desk specialist.

Observe that the default roles needed for the AI worker to perform any action aren't copied over. The user only takes the roles from the assignment group.

1. Change the 'glide.ais.query.search_operator' property to 'OR'.
2. Perform a search on Service Portal which returns more than 12 results.
3. Select the Next page arrow button.

Observe that there's no matches.

1. Navigate to the 'Search' page in the /esc portal.
2. Enter the search keyword, 'Camtasia for beginners'.

Notice that the results come from articles or catalogs instead of custom sources, and the result should be from the Learning Course table.

1. Open any profile.
2. In the Related List, create a new 'Result Improvement Rule'.
3. Create the rule with a label.
4. Check the Activate on all queries checkbox.
5. Select Create boost action after the rule is created.
6. Select the type Boost with field match(static).
7. Leave the Contains field empty.
8. Select Save.

   Notice that no errors occur.

9. Return back to the profile in step 1.
10. Select the Publish button.

Notice the error when publishing the profile.

1. Open any search application configuration, such as Service Portal.

   Observe that in the Related list Facets table, the 'Sort by' column is completely blank.

2. Open a record.

   Observe that the Sort by field has a count or some other field selected, similarly in the AI Search Admin Console.

3. Open any application.
4. Navigate to Facet filters.

Observe that the Sort by column is completely blank.

1. Log in to an instance.
2. Use the Native UI.
3. Pin the filter navigator (left navigation panel).
4. Set the browser zoom to 125% or 150%.
5. Perform a global search for a known existing record.

   Observe that the search returns 'No Record Found'.

6. Change the zoom to 80% and repeat the same search.

Observe that Global Search in the Native UI doesn't return results.

1. Log in to a Zurich instance as a portal admin.
2. Navigate to Employee Center: /esc.
3. Search for 'outlook' or 'email'.
4. From the typeahead results drop-down list, select the catalog item using the mouse.
5. When the results load, observe that the previously listed drop-down list/suggested results still display.

Expected behavior: The drop-down list should close on selection. The search state should reset or behave consistently. No residual search results should remain visible.

Actual behavior: The search results drop-down list remains visible after selection. The previous search list remains rendered.

Navigate to an instance with Zing.

On the home page, notice that the search bar doesn't load and an error appears in the console.

1. Open an instance with ESC portal and Now Assist in AI Search with the session language set to German.
2. Open Chrome and navigate to the AI Search page of ESC portal.
3. Open Chrome Developer Tools and turn on 'Responsive Design' view (toggle device toolbar).
4. Set the viewport to a narrow mobile width.
5. Confirm that the instance language is set to German.
6. Enter the search keyword 'übersicht fristen' in the search box.
7. Submit.

Expected behavior: The synthesized answer loading card displays the full loading status text without overlapping. The card container expands or wraps to accommodate the full text string regardless of locale or viewport width.

Actual behavior: The synthesized answer loading card overlaps its content. The loading status text is not fully visible. The card does not expand to accommodate the longer German locale status text string in the narrow viewport.

1. Navigate to an experience using the sn-search-top-results Seismic component (for example, Now Assist–enabled portal or Employee Center with AI Search configured).
2. Perform a search query that triggers a RAG/Genius Result response.
3. Simulate or wait for a timeout condition in the RAG response (for example, slow network, high latency instance, or LLM response delay exceeding the configured threshold).

   Observe that the search results fail to render after the timeout.

4. Without reloading the page, perform a second search query.

Expected behavior: Results render normally as the second request completes in time.

Actual behavior: Results still fail to render — asyncGeniusResultTimerExpired remains true and is never reset, blocking all subsequent renders.

1. Navigate to Platform Analytics Experience.
2. Create a dashboard.
3. Add a donut visualization:
   1. Datasource: sc_req_item
   2. Group by: variables
   3. Search for: Sample item

   Notice that only multi-choice and 'Select box' are available.

4. Create the same visualization in classic reporting.

Notice that there's a lot more options available.

Refer to the listed KB article for details.

1. Navigate to Workspaces > Hardware Asset Workspace > Inventory > All Stockrooms.
2. Select New to create a new stockroom.

Observe that the UI action to save the stockroom is not visible.

1. Navigate to Workspaces > Hardware Asset Workspace > Inventory > All Stockrooms.
2. Select New to create a new stockroom.

Observe that the UI action to save the stockroom is not visible.

1. Provision an instance with Engagement Messenger (EM) version 5.12.0 installed.
2. Set up OIDC.
3. Embed EM on a third-party site.
4. Try to log in to the site.

Observe that authentication fails (as expected) and the user is redirected to [servicenow_instance_url]/not_allowed.do.

1. Request an instance on Australia build.
2. Install com.sn_samp_master plugin.
3. Run any of the tests in app-sam-atf-test or app-sam-atf-recon-test or app-sam-atf-demodata-recon-test.

Observe that they all fail with the same error.

1. Create a test suite and add all active ATF tests to it.
2. Copy test suite to a new one.
3. Create separate user with admin roles.
4. Run two test suites in parallel browser sessions.
5. Monitor /stats.do?view=node Open Connections section with Utilization numbers.

Notice that ATF tests start to fail because they are not able to acquire DB connection.

1. License NAFC for an Australia instance.
2. Attempt to install.

Observe that it can't be installed because the BA app is not included as part of distribution.

1. Make sure the MID Server is validated.
2. Setup OAuth 2.0 on an instance.
3. Fetch the token using the Get OAuth Token UI action.

   Observe the MID Server's agent logs. There shouldn't be any error.

4. Manually expire the token.

Observe that there are errors with NumberFormatException in the MID Server's agent logs.

1. Initiate columnar migration for a large table (72 millions+ rows).
2. Navigate to Columnar Store Migration UI or DMJob.
3. Update the script to only run migration for this large table.
4. Start migration job.

Expected behavior: Migration should succeed.

Actual behavior: Migration fails for this table and subsequent retries fail as well.

1. Create a glide_list access_groups field on cmdb_ci (or tpp).
2. Ensure that it is on the partition. If not, migrate it.
3. Create a before-query business rule on cmdb_ci_server.
4. Add an encoded query on glide list that uses the dynamic filter option 'One of my groups'.
5. Execute DBSqlParser query: 'Select access_group from cmdb_ci_server'.

Observe that an error is thrown, such as 'ERROR: column __src140_cmdb0.u_access_groups does not exist'. It should use the partition instead of the root.

1. Ensure the user has a table in the MySQL remote database with a column of the type 'JSON', and that a connection is established with MySQL connector.
2. Navigate to Workflow Data Fabric (WDF).
3. Select a connection.
4. Create data fabric table for a table that has a JSON column.
5. Map the columns.

Observe that there is an unknown column type selected for JSON, and it is unavailable to map to.

1. Navigate to All > Manager Analytics > Voice Survey Mg2 DF.When opening the list view, observe the error message 'Decimal digits not present'.
2. In the list view, navigate to the column 'Last bonus Date' and select the three dots.
3. Navigate to Configure > List Calculations and deselect everything.
4. Reload the list view.

Notice that see the error disappears. Then, observe the error again upon re-selecting the list calculations.

Attempt to query an SN table via Trino.

Expected behavior: By default, the user should have access if they have read access to the table.

Actual behavior: The user must have specific data_fabric access granted to that table to run the query.

1. Provision a table with a binary datatype in the mySQL database.
2. Create a DF table.

Notice that the column with type binary is greyed out and cannot be mapped.

1. Navigate to DMC and select any table with rules in archive rule active.
2. Select the table and navigate to a 'Rule detail' page, change the conditions and select Update.

Expected behavior: The rule should change to an inactive state, and a message should be displayed indicating that the rule was made inactive due to condition changes.

Actual behavior: Active button in create flow should be inactive when the rule is changed.

1. Navigate to the Create Rule module.
2. Create a rule with conditions and save it successfully.
3. From the rules list, select the Edit option for the created rule.
4. Wait for the 'Edit rule' page to load.
5. Observe the loading behavior of the conditions section.

Expected behavior: The rule edit page should load within a few seconds. All configured conditions should be displayed correctly and consistently every time. Add a loading indicator so the user is aware before taking any action or updating an incomplete record.

Actual behavior: The 'Edit rule' page takes around ten minutes to load the conditions. In some cases, the conditions do not load at all, leaving the edit page incomplete or blank.

Activate com.glide.db.columnar.archive to enable offloading.

Expected behavior: DMC should show up for the tier2 storage tile.

Actual behavior: DMC does not show up for the tier2 storage tile.

1. Create a analytics source that collects data 'Daily' on Problem table with timezone America/Los_Angeles and create DS indicator on it with no conditions.
2. Run the collection jobs.
3. Navigate to Source state table(pa_dm_source_state) and set start_date_time as 30 days ago from current date and set last_delta_time as current date.
4. Navigate to pa_cdc_fact_problem and Change start_date_time to ten days of current date, so that scores before that day is 0.
5. Navigate to KPI details for the newly created indicator, select the 11th day from current day where the zero score starts and select Show Records.

Notice that Show records shows scores one day ahead.

1. Check the ACL and Java code for usage of the snc_required_script_writer_permission role and conditional script writer group.
2. Turn the switch on or off.

Observe that the feature may break.

1. Provision an instance with com.glide.dsb.cc installed on the source instance.
2. Install and setup DSBs on the target instance.
3. Trigger a clone from the source to target.
4. Wait for the clone to finish.
5. Run 'gs.info(new sn_dsb.API().isEnabled())' on the target instance

   Notice that DSBs are disabled.

6. Check the list of DSBs on the target instance

Notice that none of the DSBs have been rebuilt.

1. Create a Zurich instance.
2. Enable sandboxes.
3. Create sandboxes.
4. Upgrade instance to Australia.
5. After upgrade has completed, check the sandbox states.

   Notice that they're all stuck in 'Queued'.

6. Verify that events are stuck in a 'Ready' state in a sysevent (filter by name = dsb.create).

Check sys_trigger for 'Process Sandbox Operation Queue' - there should be one per controller, but there are none.

Refer to the listed KB article for details.

1. Configure the 'Email' related tab on the Workspace UI Builder view of sn_hr_core_case_workforce_admin.
2. Navigate to the HR Agent Workspace.
3. Open any HR record by impersonating an HR Admin.
4. Send an email using Compose Email.
5. Select a user who has two user profiles (under 'To:'). One active and one inactive.
6. Navigate to the 'Emails' related tab.
7. Open any email record.

Expected behavior: Each recipient should be displayed only once in the notification, without any duplication.

Actual behavior: Duplicate recipients are displayed in the notification for the same email under the 'Emails' related tab.

1. Assign 'notification_admin' role to 'abel.tuter'.
2. Impersonate 'abel.tuter' and navigate to 'sysevent_email_action_list.do'.
3. Add 'Weight' column to the list view.
4. Add a filter 'weight>10'.

Expected behavior: The user should be able to perform query_range operation and view the results.

Actual behavior: The user observes the error 'Part of the query on sysevent_email_action has been ignored because of insufficient access for 'query_range' operation on sysevent_email_action.weight'.

Refer to the listed KB article for details.

1. Create an employee case for any user with the HR service as PIP.
2. Move the case to the 'ready' state.
3. Create another case with same service and try to submit it.

Expected behavior: The case shouldn't be created and an error should be thrown.

Actual behavior: It gets submitted and a new case is created.

1. Create topology.
2. Stop the 'Grouping' job.
3. Send alerts on CI1 and CI8.
4. Run the 'Grouping' job.

   Observe that in em_ci_graph_reuse, the graph of CI1 also contains CI6, CI7, Ci12, etc.

5. Close the group.
6. Send alert on CI6.
7. Run the 'Grouping' job.

   Observe that the alert is added to the staging table but no group is created.

8. Send alert on CI8.
9. Run the 'Grouping' job.

Observe that CI8 and CI6 create a new CMDB group. However, alerts should not be grouped.

1. Open a Zurich instance.
2. Run Service Analytics group alerts using RCA/Alert Aggregation.

Observe that it can get stuck under concurrent load.

1. Send a SNMP V1 trap.
2. Navigate to the em_event table.

See that there are no events with a node.

1. Log in to any Australia instance and assign abel.tuter role 'flow_admin'.
2. Impersonate abel.tuter and navigate to sys_flow_log.
3. Perform any range query operation ('contains') on the field Context.

Expected behavior: abel.tuter with role 'flow_admin' should be able to perform query_range operation and view the results.

Actual behavior: Part of the query on sys_flow_log has been ignored because of insufficient access for 'query_range' operation on sys_flow_log.context.

1. Give a user flow_admin.
2. Navigate to the list for sys_hub_flow.
3. Try to query internal_name Starts With.

Notice that the list blocks that query.

Refer to the listed KB article for details.

1. Log in to any Zurich instance.
2. Activate the ER Interview Summarization Skill.
3. Open any ER Interview.
4. Select Summarize.

Observe that the Copy to Interview Summary button isn't visible after the summary is generated.

1. Create or open a Map / Data Transfer Action.
2. Add duplicate target fields in the target section.
3. Attempt to save or publish the record.

   Observe that an error is shown.

4. Check the database tables.

   Observe that partial target data is saved in sys_rte_eb_entity in JSON and no corresponding record is created in sys_rte_eb_definition.

5. Fix the duplicate target field name to make it unique.
6. Attempt to save or publish again.

Observe that the record is still not inserted properly.

Add the data mapper step and create a step input and then map some source fields to target fields.

Notice that when the pill is dropped for the first time it doesn't render the source tree.

1. Open Flow Designer and create a new action.
2. Add an Input variable with a label containing a space. For example, 'First Name' (internal name auto-set to 'first_name').
3. Add a Map and Transform step to the action.
4. Configure the Map and Transform step to map the First Name input field to an output or target field.
5. Save and publish the action.
6. Trigger the action with a test value for First Name.
7. Observe the output.

Expected behavior: The RTE correctly identifies the source field by its internal name (with underscores) and transforms the data as configured.

Actual behavior: The RTE skips source fields whose labels contain spaces, resulting in empty/null output for those fields.

1. Perform a cross region Move or Clone on Washington DC instances (with or without EKMS, seeing same behavior).
2. The Rekey is complete but the source IRK record is still present in the target sys_kmf_external_key table.

Expected behavior: The source IRK should be deleted and Grant rule is revoked.

Actual behavior: Observe that the grant rule is not revoked.

Refer to the listed KB article for details.

1. Open an published KB article.
2. Navigate to the related list.
3. Find the 'Related articles' tab.
4. Select the Edit button.
5. Add another KB article as a related list.

Observe that from the related list 'mapping record' achieved on step 4, users can edit the mapping record to and add the same record as the related article.

1. Navigate to Knowledge Management > Import Article
2. Import a Word document.
3. Open the created Knowledge Article.
4. View the source HTML.

Observe that the output contains flexible/relative styling and there's no static absolute positioning. Compare against a Knowledge Article created via the GRC/IRM policy flow and note the static positioning difference

1. Open CSM/FSM Workspace.
2. Navigate to My Lists.
3. Open a list from either 'Created by me' or 'Shared with me'.
4. Select Save a copy.
5. Open the new copy.
6. Share the copied list with a user or group.

Expected behavior: After sharing the copied list, the Share icon appears immediately without requiring a page refresh.

Actual behavior: After sharing the copied list, the Share icon doesn't appear immediately in the 'My Lists' overview page. It only becomes visible after refreshing/reloading the page.

1. Open any record list in a Zurich instance.
2. Perform grouping on any field.
3. Selec the Show all group header action.
4. Select the Back button.

Observe a JS exception in the console, preventing the Back button action.

1. Create a Boolean Data Type metric.
2. Run the script in scripts background.

Expected behavior: Valid query result.

Actual behavior: An error occurs, 'Script execution error: Script Identifier: null.null.script, Error Description: com.snc.clotho.ClothoException: com.glide.db.GlideSQLException: Element doesn't exist. Table: mb_dex_installed_app_metrics col: false.'

Refer to the listed KB article for details.

1. In any instance, activate the Field Service Management Demo plugin.
2. As admin, double check that the List screen 'My tasks' has the default filter.
3. Sign in as a user from the Agent app.
4. Edit a task or pick a task that has an expected start date that is easy to filter.
5. Select My tasks > See all.
6. Check that online filtering on 'Expected start' works correctly and one task is picked.
7. Download offline.
8. Go offline.
9. Perform the same filtering in offline mode.

Expected behavior: One task is picked and displayed after filtering on 'Expected start'.

Actual behavior: All tasks are displayed after applying the filter.

1. Provision an instance with Field Service Management with demo data installed.
2. Log in to the agent app with a field agent user.
3. Navigate to the offline settings.
4. Download the offline cache.
5. Go offline.
6. Open a work order task form screen and make an edit.
7. Open a separate work order task and make an edit.
8. Navigate to settings.
9. Sync online.

Expected behavior: In the offline synchronize response, the Document > Data > Tables for each execution result should only contain the record that was edited.

Actual behavior: The Document > Data for each execution result contains the data for both records in each execution result.

1. Log in into any Zurich instance with Now Assist capabilities enabled.
2. Impersonate any 'ITIL' or 'sn_hr_core.admin' user.
3. Open any workspace (for example, SOW or HR agent workspace) and open an active Incident or HR record.
4. Change the Assigned to field to a different user.
5. Open now assist panel and 'pin' it.
6. Select the Assign to me button.
7. Select View Plan.

Expected behavior: Plan is displayed on the NAP.

Actual behavior: The NAP is blank.

1. Type 'What is spam?'.
2. Select any feedback thumbs up/down icon immediately.

   Observe that feedback is successfully stored in the Gen AI log table.

3. Deselect submitted feedback or submit different feedback.

Expected behavior: Feedback should be updated in the Gen AI log table.

Actual behavior: Feedback isn't updated in Gen AI log table.

1. Run any GenAI Skill from background.
2. Update/create any record immediately after the GenAI Skill call.

Observe that the record is incorrectly marked as 'Updated by Now Assist'.

1. Create an OAuth integration with any flow (for example, auth code or client credentials).
2. Create a REST Message V2 record.
3. Set 'Authentication type' to 'OAuth 2.0'.
4. Set 'OAuth profile' to the OAuth entity profile created in the first step.
5. Attempt to set the MID Server.

Expected behavior: The MID Server can be selected, enabling the OAuth credential to be sent as an authorization header via MID.

Actual behavior: The MID Server can't be selected. OAuth credentials can't be sent as an authorization header via MID.

1. Create a pie chart grouped by state from Visualization Designer.
2. Invoke /api/now/sg/visualization.
3. Make sure that the color code is returned for incident state eight in the response.
4. Navigate to the sys_report_chart_color table.
5. Filter table:incident and element:state.
6. Update the color code for incident state 8.
   1. Color name: Use the search box and choose alert--info-1.
   2. Color: Open the alert--info-1 record using the i preview icon and copy the color code from there (#5EABDF).
7. Save the record.
8. Invoke /api/now/sg/visualization again.
9. Check the color code returned for incident state 8.

Expected behavior: The color code should be returned as #5EABDF for incident state 8.

Actual behavior: The DB update isn't reflected in the API response as the cache still has the old color code value.

Refer to the listed KB article for details.

1. Create a dashboard.
2. Share the dashboard to the role 'dashboard_user' with view mode.
3. Impersonate an ITIL user.
4. Open the dashboard.

Expected behavior: The dashboard should be visible to the ITIL user.

Actual behavior: The dashboard is not found.

1. Create a dashboard in Next Experience.
2. Add a Rich Text component.
3. Add content with embedded styles.
4. Save the dashboard.
5. Reload the dashboard.

Observe that text is displayed without styling.

1. Provision an Australia or Zurich instance with a language plugin installed.
2. Switch the language to the installed one.
3. Create a dashboard.
4. Add two new tabs.
5. Give a name to both.
6. Save the dashboard.
7. Switch to English.
8. Rename both tabs.
9. Save the dashboard.
10. Switch back to the initial language.

Expected behavior: The tab names should be preserved and not overridden.

Actual behavior: The tabs now contain English names instead of the ones for the specific language.

1. Create a new PA dashboard.
2. Add a multi-select filter with source User > Name.
3. Select Configure values for filter source and uncheck all names starting with A, B and C and save.
4. Save and exit edit mode.
5. Select the filter.

Expected behavior: Unchecked values should be filtered out for filter values.

Actual behavior: All values for User.Name (including the ones starting with A, B and C) are loaded and available for selection.

1. Create a Core UI dashboard in the global domain.
2. Migrate the dashboard partially in the global domain.
3. Switch to the ACME domain open-migrated Next Experience dashboard.
4. Rollback the dashboard inside the ACME domain.
5. After rollback, it creates a Core UI dashboard in the ACME domain with an overrides value to the original dashboard.

Observe that the Core UI dashboard isn't accessible from the ACME domain after rollback. Also, when users migrate a global Core UI dashboard inside a domain, then duplicate bridge records are generated.

1. Create a CoreUI dashboard.
2. Add a report widget that's used in other dashboards and set the header alignment to be center.
3. Migrate the dashbooard.
4. Verify that the par_visualization record is created for the sys_report.
5. Open the dashboard in PAE.

Expected behavior: The header alignment settings are migrated.

Actual behavior: The header displays with the default settings.

Refer to the listed KB article for details.

1. Open a Playbook with a questionnaire activity (or any activity with only one form section that has a heading defined).
2. Execute the Playbook to reach the questionnaire activity.

Observe that the form section heading (for example, 'Variables') is displayed.

1. Have a parent record on a workspace with a work notes field.
2. Have a playbook activity with a form that contains the Work notes field.
3. Submit the form via Update or Mark Complete.

Expected behavior: The work notes component does not get the work note data from the playbook form.

Actual behavior: The work notes component gets the submitted work note data.

1. Navigate to process models.
2. Take any playbook project and note the average duration on the project card.
3. Navigate to Workbench and note the average duration shown in the header.
4. Open Global Metrics and note the average duration.

Expected behavior: On the project card, average duration should retrieve the mean duration values which has been fixed and is being called from the front-end side.

Actual behavior: On the project card, the minimum duration value is sent for average duration. Therefor, there is a discrepancy in the stats shown for average duration.

1. Open any record hierarchy (for example, manager hierarchy on sys_user).
2. Wait for the associated sys_record_hierarchy_status record to reach a READY state.
3. Delete the associated sys_record_hierarchy_status record.
4. Clear the cache by visiting /cache.do.
5. Visit the associated sys_record_hierarchy record.
6. Attempt to perform a reconciliation.

Notice that a message appears saying that reconciliation has been scheduled, but no sys_record_hierarchy_status record is created and no event is posted to sysevent.

Refer to the listed KB article for details.

1. Create an update set on dev.
2. Export as XML.
3. Import, preview, and commit the update set on test without using ReleaseOps.

   Observe that the update set is committed without an info record and the controller isn't aware of the commit.

4. Promote the update set to test using ReleaseOps.

Expected behavior: The deployment request fails.

Actual behavior: The deployment request doesn't fail and the update set gets committed twice.

1. Provision an instance with the 'Domain Separation' plugin and demo data installed.
2. Create a user in a subdomain.
3. Assign a few roles (snc_internal and another role) to the user.
4. Check the sys_user_has_role records.

   Notice that a new record with the 'snc_required_script_writer_permission' role is created with an empty domain path. Make sure to open as a standalone list, because the records aren't visible in the related list.

5. Open the sys_user_grmember table in the list view.

   Notice that the user is added to the 'Conditional Script Writer' user group. Again, it's not visible under the Groups related list because the domain path is empty.

6. Disable the user (active=false).
7. Reenable the user (active=true).
8. Check the sys_user_has_role list.

Observe that a new record with the snc_required_script_writer_permission role is created. The domain path is again empty/incorrect.

1. Create a scope and a new table.
2. Create a relationship (sys_relationship) record with the following configuration:
   1. Name: Created by
   2. Advanced: true
   3. Applies to table: New Scoped Table
   4. Apply to: gs.debug('Table Name: ' + table_name);
3. Open the New Scoped Table form.
4. Right-click and navigate to Configure > Related Lists.

Notice that when checking the system logs, the following error message occurs: 'com.glide.script.RhinoEcmaError: 'table_name' is not defined. : Line(1) column(0) ==> 1: gs.debug('Table Name: ' + table_name); Stack trace: at :1.'

1. Navigate to SP portal.
2. Select Catalog.
3. Select any of the categories on the left.

Observe that the catalog items don't appear. The URL keeps changing but stays on the same page.

Refer to the listed KB article for details.

1. Create Manual Service.
2. Add relations to CMDB where the parent is 'Service' and the child is some CIs.
3. Add duplicated records of one relation using the BG script to the cmdb_rel_ci table.
4. Add one more relation (not duplicated) for the service.
5. Select Update with changes from CMDB.

The entry point for the latest added CIs won't be added to the map.

1. Take any fluent app that has unpublished wf_workflow_version records.
2. Run now-sdk download/transform.

Notice that the command fails with the 'Unexpected error' log, and there is an exception logged on the instance log.

Scenario 1:

1. Make sure an app exists on the instance with Form metadata containing a couple sections and elements.
2. Add or update elements in the UI section.

   Notice that the sys_update_xml table updates with changes.

3. Perform download using FluentDownloader.
4. Provide the last updated timestamp.

Expected behavior: All data from the last updated timestamp shown in sys_update_xml gets downloaded.

Actual behavior: Form sections are not downloaded. Scenario 2:

1. Make sure an app exists on the instance with Form metadata containing a couple sections and elements.
2. Add or move existing sections in Form.

   Notice that sys_update_xml updates with changes.

3. Perform download using FluentDownloader.
4. Provide the last updated timestamp.

Expected behavior: All data from the last updated timestamp shown in sys_update_xml gets downloaded.

Actual behavior: Form isn't downloaded. Only the UI section gets downloaded.

1. On the latest base instance version, create a client script on the incident table with the UI type 'Mobile/Service Portal'.
2. Set the type to onLoad.
3. In the script, add a line with g_form.addDecoration, such as: g_form.addDecoration('description', 'icon-star', 'Please provide steps to reproduce the error.');.

Expected behavior: Field decorations display a star symbol on the Description field.

Actual behavior: Field decorations don't appear.

1. Create a record producer.
2. Create three date variables.
3. Make one of the variable 'hidden'.
4. Open the record producer in the portal.
5. Hover over the Date field.

Expected behavior: A tooltip is visible.

Actual behavior: A tooltip isn't visible.

1. Set up a direct Kafka cluster.
2. Create a Kafka Flow Trigger Consumer with a topic alias, publish it.
3. Make sure the flow reporting setting is set to 'Full' or 'Basic'.
4. Push a message on Kafka so this flow consumes it.

Notice the error in execution log.

1. Make a daisy chain of discovery schedules to trigger one after another.
2. Set 'Even if canceled' to 'true'.
3. Run discovery and cancel it.

Notice that even though the next discovery should be triggered, it does not always happen.

Open a syslog_transaction record in form view.

Observe that the http_response_code field may be editable.

1. Open an instance with the 'Domain separation' plugin installed.
2. Create a technical dashboard within the global domain.
3. Add a data visualization to it.
4. Switch to another domain.
5. Create a page override.
6. Add an additional data visualization to the overridden page.
7. Preview the created dashboard pages from UIB by switching domains.

   The page content changes as expected.

8. Open the technical dashboard via the Platform Analytics experience and switch domains to check the page content.

Expected behavior: The overridden page variant should be displayed for the domain.

Actual behavior: Only the global page variant is displayed in the domain.

1. Launch JAWS.
2. Impersonate Beth Anglin.
3. From the 'All' menu, navigate to the certificate management page.
4. Navigate the page to find tabs like 'Tasks', 'Records', and 'Download'.

   Note that the currently selected tab does not have a selected status reported by JAWS.

5. Select a tab.

   Note that tab now reports its selected status.

6. Reload the page, and note the lack of selected status.

Expected behavior: The selected status of currently selected tab is always reported.

Actual behavior: The selected status of the tab is not reported on first page load.

1. Navigate to a base instance with Hardware Asset Workspace (plugin: sn_hamp).
2. Navigate to All > Hardware Asset Workspace.
3. Open at least one other record.
4. Set window size to 1280x1024.
5. Zoom the window in to 400%.

   The second tab in the tab container is now in a More drop-down list.

6. Expand the tab.

Expected Behavior: Content does not require scrolling in two dimensions (both vertically and horizontally).

Actual Behavior: Content requires scrolling in two dimensions (both vertically and horizontally).

1. Start a node with a set system_id.
2. Note the node_id.
3. Modify the system ID to have 'AMS' capitalized.
4. Restart the instance.
5. Observe that the node ID is now different.
6. Modify the system ID back to the original value.
7. Restart the instance.

Observe that the node ID is now back to the original.

1. Set up an instance with Now Assist and AI Search.
2. Perform a search that returns multiple sources.
3. Close the conversation.

Expected behavior: interaction.transcript should show the sources that are displayed to the chat request.

Actual behavior: interaction.transcript shows all sources in the smartLinksData regardless of whether or not they are shown to the user.

1. Invoke an Agent that invokes a topic-as-a-tool.
2. Ensure that the topic has a Boolean/yes/no question, such as 'Do you like Fridays?'.
3. When the runs and asks if the user likes Fridays, answer with 'What is spam'?
4. Verify that the topic stops and there's an LLM answer about spam.
5. Perform a Q&A about spam.
6. Type another phrase that would trigger an agent, where the 2nd agent is configured with a tool that is-a topic.

Notice that the second topic doesn't run correctly. The conversation is stuck at this point.

1. Create a LLM topic.
2. use vaContext.url in the script.
3. Call the LLM topic through the nextWave NAP.

Notice that the vaContext.url is undefined.

1. Log in to an instance.
2. Open AI agent playground.
3. Select the Demo Next Best Action Agent agent and run it.

   Observe that the execution gets stuck right after the first tool call.

4. Check the sys_generative_ai_log for the above conversation.

Observe that there are two LLM calls.

1. Navigate to any instance that has NextWave.
2. On the NextWave home page, verify it displays the main search bar page.

The search placeholder should be 'Ask Otto or search anything' for the AICT assistant.

1. Navigate to Assistant Designer for NAP Platform assistant
2. Navigate to 'Display experiences' page

Notice that display experience is set to 'Enhanced' but if instance has nextwave setup, an Edit button will be available next to it allowing user to change display experience to 'Premium'.

1. Navigate to /esc and open the chat bot.
2. Type 'Presentation device' manually and start the conversation.
3. Select options or enter the asked information and proceed with the flow.

Observe the incorrect highlighted message/strings: '1 選択済み' and '送信 1' - ' 1 Selected' and 'Send 1'.

1. Start a new chat.
2. Give some input.
3. Start another chat before the response comes in.
4. Give a new input in the new chat.

Notice the response from the previous chat sometimes show up in the new chat, and when refreshing the page, it's all in the same conversation, and the new input entered is merged in the previous conversation.