Washington DC Patch 7

  • Release version: Washingtondc
  • Updated September 5, 2024
  • 49 minutes to read
  • The Washington DC Patch 7 release contains important problem fixes.

    Washington DC Patch 7 was released on September 5, 2024.
    • Build date: 08-31-2024_1809
    • Build tag: glide-washingtondc-12-20-2023__patch7-08-21-2024
    Important:
    For more information about how to upgrade an instance, see ServiceNow Upgrades.

    For more information about the release cycle, see the ServiceNow Release Cycle.

    Note:
    This ServiceNow AI Platform major family release is now available in ServiceNow's Regulated Market environments. For more information about services available in isolated environments, see KB0743854.

    For a downloadable, sortable version of the fixed problems in this release, click here.

    Overview

    Washington DC Patch 7 includes 175 problem fixes in various categories. The chart below shows the top 10 problem categories included in this patch.

    Figure 1. Top 10 problem categories
    Fixed issues grouped by problem categories bar chart

    Changes in Washington DC Patch 7

    Enable Now Assist experience in Virtual Agent API
    Be sure you are on Washington DC Patch 7 or above. If you are on an earlier version, see KB1699602 article in the Now Support Knowledge Base.

    Notable fixes

    The following problems and their fixes are ordered by potential impact to customers, starting with the most significant fixes.

    Problem Short description Description Steps to reproduce

    AI Search

    PRB1778235

    There's an 'Invalid or unexpected token' error message on a browser console when accessing any page on a portal after upgrading to a Washington DC version The issue can be seen on any portal, like 'sp' or 'esc', as well as an empty testing page. There shouldn't be an error message on the browser console.

    Database Persistence - Data Access

    PRB1717895

    KB1629327

    The 'OR' logical operator in a related list condition query behaves as the 'AND' operator When reporting over a table, using a related list condition with the 'OR' logical operator, it seems to have the same effect as using 'AND'.

    Refer to the listed KB article for details.

    Legacy Agent Workspace

    PRB1722642

    When two users open an HR Case in HR Agent Workspace, the second user can't see that the first is viewing When user X opens a HR Case on the HR Agent Workspace, and then user Y opens the same HR Case on the HR Agent Workspace, user X can see that user Y is viewing the HR Case, but user Y can't see that user X is viewing the HR Case.
    1. Provision an instance with the 'sn_hr_ws' and 'sn_hr_agent_ws' plugins on a Vancouver instance.
    2. Log on to the Vancouver instance.
    3. Impersonate a user.
    4. Create an HR Case.
    5. Navigate to the HR Case on the HR Agent Workspace.
    6. Open an incognito tab on the browser.
    7. Log on to the same Vancouver instance.
    8. Impersonate a different user.
    9. Navigate to the same HR Case on the HR Agent Workspace.

    Observe that the first user can see that the second user is viewing the HR Case, but the second user can't see the first user is viewing the HR Case.

    Software Asset Management

    PRB1638689

    KB1227735

    The business rule 'Bookkeeping for dedup on delete' doesn't check the normalization status of installs

    Refer to the listed KB article for details.

    UI Form Administration

    PRB1665993

    g_aw.closeRecord() doesn't work for new records in a workspace After creating a Save and close UI action, it isn't possible to close the current record tab for new records.

    All other fixes

    Problem Short description Description Steps to reproduce

    Activity Stream

    PRB1709271

    The user is unable to select email hyperlinks in the activity stream of Service Operations Workspace (SOW) There's a degraded user experience for using hyperlinks in email. They may intermittently not open a new tab with a URL address.
    1. Log in to a Vancouver instance.
    2. Navigate to Service Operations Workspace (SOW) > Create or view an incident record.
    3. Ensure that an email is associated to the incident record.
    4. View the activity stream of the incident record.
    5. Select the hyperlink for the email.

    Expected behavior: The system should load the new tab with the URL address.

    Actual behavior: The system intermittently doesn't load in a prompt manner.

    Activity Stream

    PRB1763824

    A space/line break is observed in UI15's activity stream for emails The issue is reproducible only in Washington DC. It occurs only when the activity formatter is set to the 'UI15' view through the 'use.concourse' user preference.
    1. In any Washington DC instance, navigate to sys_user_preference.
    2. Find the 'use.concourse' system-wide preference.
    3. Set the value of this preference to false so the activity streams are rendered as UI15.
    4. Navigate to any incident record.
    5. Send an email from the record.
    6. Navigate to the sys_email table.
    7. Set this email as 'sent' so it displays in the activity stream.

    Observe that in the UI15 activity stream where the email is displayed, there's an unexpected space/line break.

    Activity Stream

    PRB1763981

    Activity streams need a maximum length to guard against tables with excessive audit relations Configuration Item (CI) records with large amounts of related sys_activity records load slowly in a configurable workspace.
    1. Add a large number of incidents, 10 K - 100 K.
    2. For each incident, on the incident form in Service Operations Workspace, set the Service (business_service) field in the 'Impact' section of the form to the 'Email' service.
    3. Ensure that the 'Email' service of the CI activity stream is removed; 'GlideActivity.removeStream' in sys.scripts.do.
    4. Open the 'Email' service CI by selecting the 'Open Record' icon (info) in the Service field in the 'Impact' section of the form.

    Expected behavior: The CI opens quickly.

    Actual behavior: The CI gets stuck loading and there's statements in the log file about handling an extremely large result set.

    AI Search for Virtual Agent

    PRB1789779

    Users are seeing a 'bad content' error for 'what is spam?'
    1. Set up a large language model conversation on esc.
    2. Log in to the esc portal.
    3. Open the Web Client.
    4. Type 'what is spam?'

    Observe the 'bad content' error.

    AI Search

    PRB1720477

    Translated Field Helper generates unexpected AI Search documents for installed languages Translated Field Helper generates unexpected AI Search documents by searching preferred_language of sys_choice table for installed languages of the sys_user.
    1. Provision an instance without the i18n pack plugin.
    2. Navigate to sys_choice.
    3. Create entries on the sys_user table in the 'preferred_language' column.
    4. Ingest a sc_cat_item record with language.

    Observe that two documents are generated for one glide record.

    AI Search

    PRB1779119

    After upgrading to Washington DC, the user sees a red banner ACL message when viewing incidents or case tasks The user observes an error with the banner, 'Part of the query on live_message has been ignored because of read security rules on live_message.chat_message', or 'Part of the query on live_message has been ignored because of read security rules on live_message.state'.

    AI Search

    PRB1780839

    Incorrect values are logged for triggered Genius Results (GR) in an async case When the async mode is 'GR_ONLY', 'GR Placeholders' are returned for every async GR that is executing when returning the initial search response. These placeholders end up in the searchAnalyticsPayload on that response, which is used to log the initial SEARCH_EVENT. As part of the logging of that SEARCH_EVENT signal, those placeholder GRs are then logged to the triggered GRs table. However, there's no guarantee any of them actually resulting in valid GRs, and if they do, there's no guarantee they were shown to the user.

    AI Search

    PRB1785294

    Suggestions for recently viewed results (PERSONAL_CLICK suggestions) with external URLs don't navigate correctly in portal
    1. Open Service Portal on an AI Search-enabled instance.
    2. Search for 'Change Password'.
    3. Select the regular (non-Genius) search result for 'Change Password'.

      Observe it brings the user to a page outside of the portal (e.g. $pwd_change_error.do).

    4. Run the scheduled job to process the queued click signals, to get 'Recently Viewed' suggestions.
    5. Return to the portal home or search page, and type 'Change' in the search box.
    6. Select the 'Recently Viewed' suggestion that says 'Catalog Item' underneath it.

    Expected behavior: The user should be taken to the same URL used when selecting the regular search result.

    Actual behavior: The user is taken to the generic portal form page.

    AI Search

    PRB1786056

    The carousel Genius Result position isn't added to the signal
    1. Navigate to the Next Experience homepage.
    2. Search for a term that triggers a carousel with more than one result.
    3. Provide feedback or do an action like a request for one of the items from the multiple items on the carousel.

    The carousel Genius Result position should be added to the signal, but isn't.

    AI Search

    PRB1788062

    Search is down for some time when upgrading from Vancouver to Xanadu An error is seen: 'AISearchService: Error is thrown from searchQueryService.search() : : java.lang.NullPointerException: Cannot invoke "com.glide.db.meta.Table.getTableIterator()" because "result" is null: com.glide.ais.configuration. Datasource.loadSemantic SearchMappings(Datasource.java:351) com.glide.ais. configuration.Datasource. loadFromDB(Datasource.java:250)'.
    1. On a Vancouver instance, provision an instance with Now Assist ITSM installed.
    2. Attach Now Assist actions and Now Assist Q&A to the Service Portal search profile.
    3. Perform queries on the portal.
    4. Upgrade Glide to Xanadu.

    Notice that the regular search is down during the upgrade.

    AI Search

    PRB1788078

    sys_search_event doesn't include grConfig in gr_results for Genius Result feedback and clicks
    1. Select Feedback or the Genius Result action.
    2. Wait for the search event to be flushed from the sys_signal_event_queue.

    Observe that the entry in sys_search_event.gr_results has an empty 'grConfig' value.

    AI Search

    PRB1793717

    The Natural Language Query (NLQ) 'Switch' table and 'Show' table doesn't emit feedback from sn-search-top-results

    In Washington DC, select any non-feedback button on an NLQ result.

    Observe that no feedback is returned through the network calls tab, and that no genius result event was logged in the search's sys_search_event.

    Analytics Data API

    PRB1788888

    Data broker calls fail after an upgrade The Analytics Data API needs to sanitize the request coming from data brokers, and remove placeholder values.
    1. Create a UI Builder (UIB) page.
    2. Add a data visualization data broker (table).
    3. Fill out the properties of the broker so that the response is a correct data visualization response.
    4. Save the page.
    5. Navigate to the page definition.
    6. In the Data field, remove the useDataCache property value from the data broker.
    7. Navigate back to UIB and refresh the page.
    8. Open the data broker.

    Expected behavior: The response is the same as in step 3.

    Actual behavior: There's an error.

    Antivirus Scanning

    PRB1734157

    Update com.glide.snap.url to snap-v2 for the release of Glide The virus scan isn't working when attaching a file to the File Attachment field.

    Application Install Engine

    PRB1743934

    Activating plugins using the new application manager logs 'system', where the previous manager logs the user who activated the plugin When installing a plugin using the new application manager, it logs 'system' in the sys_plugin_log. This makes it hard to audit who installed the plugin on the instance, where the previous manager logs the user who activated the plugin, the audit was easier.

    Application Install Engine

    PRB1786797

    Batch Install doesn't handle circular dependencies for non-app V2 apps The BatchInstallDependencyMap.java class doesn't handle circular dependencies when the app being installed isn't a V2 app. The shouldEvaluateDependencies method incorrectly returns true when a circular dependency is encountered. When loading a non-V2 app and there's a dependency pair that's already seen, shouldEvaluateDependencies returns true because it checks the list of V2 plugins visited, which is empty because it's not a V2 app.
    1. Open a Washington DC instance.
    2. Attempt to batch install App Engine Studio.

    Observe that the progress bar doesn't go beyond 0%, and eventually a StackOverflow message appears.

    Application Install Engine

    PRB1786807

    Conditional applications weren't installed via new app management
    1. Procure a new Washington DC instance.
    2. Navigate to Store.
    3. Request the 'Healthcare and Life Sciences Service Management Core' application.
    4. Complete the installation of 'Healthcare and Life Sciences Service Management Core' on the instance via system applications.
    5. Navigate to system applications.
    6. Verify sn_cwf_wrkspc install's status.

    Expected behavior: sn_cwf_wrkspc should be installed or CSM/FSM Configurable Workspace is available in the instance.

    Actual behavior: sn_cwf_wrkspc isn't installed or CSM/FSM Configurable Workspace isn't accessible in the instance.

    Application Manager

    PRB1780803

    KB1649374

    Remote app table entries are populated with 'undefined' in the 'sys_code' column Installer copies sys_code values from sys_remote_app table to sys_store_app table and used it for all app installations. However, it didn't account for records which have 'undefined' legacy values. As a result, the existing records with 'undefined' values causes failure to load conditional content in the plugin.

    Refer to the listed KB article for details.

    Appointment Booking

    PRB1789260

    Personal events on a technician calendar are ignored when booking an appointment for an appointment booking user There's also an error in the CSM/CSP portal when booking/rescheduling an appointment as an appointment booking user.
    1. Create a personal event for an agent.
    2. Ensure that only that agent is available in the group, so that DS checks availability and events for that user alone.
    3. Log in as a user in the CSM portal.
    4. Book an appointment.

    Expected behavior: An appointment slot shouldn't be visible when an agent is on timeoff, or when booked for an appointment for any other available slot. The appointment should be booked with no error thrown.

    Actual behavior: Appointment booking/reschedule isn't honoring the time where a technician is on timeoff and displays slots. Also, users are seeing an error message on the screen.

    Asset Management

    PRB1778297

    KB1649341

    The Hardware Asset Management (HAM) 'Incident' view rule within Asset Management Workspace targets Agent Workspace The Workspace field is set to 'Agent Workspace'. The HAM 'Incident' view rule shouldn't target Agent Workspace. The Agent Workspace field should remain empty.

    Refer to the listed KB article for details.

    Asset Management

    PRB1781337

    The 'Stock Runner' job is failing with an error The 'Stock Runner' job is failing with an error: 'ProcessStockRules: ReferenceError: "sn_eam" is not defined.: no thrown error.'

    Automated Test Framework (ATF)

    PRB1764100

    ATF steps that execute without explicitly setting 'Success' or 'Failed' statuses can leave pending suite results and stuck cloud run executions Even after all tests are finished, the suite and progress tracker both stay in a 'Running' state. The cloud runner stays active even though there's nothing left to run, until it eventually times out much later. This can happen any time an ATF test step doesn't get set to a 'finalized' state (i.e. Success/Failed/Error), including because the step author simply forgot to do it, or because an error was thrown.

    Automated Test Framework (ATF)

    PRB1769587

    ATF scheduled client test runners can sometimes get unimpersonation errors This can cause other tests in the same test runner to also have unimpersonation errors.

    Case and Knowledge Management for HR Service Delivery

    PRB1785629

    KB1698772

    The 'Manage Human Resource' catalog page isn't loading An error displays in the console: 'Uncaught SyntaxError: Unexpected token ')...'

    Refer to the listed KB article for details.

    Change Management

    PRB1790024

    A review flow should only attach one post-implementation review task Review flow's configuration needs to be corrected. Because the creation of the task doesn't mean the flow waits until the task is complete, it creates the task and then the flow ends. Based on the trigger condition of the flow, this means that when a user updates the record while it's in the review state, it keeps triggering the flow and creating new tasks.
    1. Make an unauthorized change model available in 'Create New'.
    2. Create an unauthorized change.

      A single post-implementation review task is associated.

    3. Update the change request's short description.

    Another task is associated because another flow was attached.

    Field Encryption

    PRB1781726

    KB1648610

    Duplicate Encryption Configuration field records for tables in the CMDB hierarchy cause errors when the tables are queried Duplicate Encryption Configuration field records for tables in the CMDB hierarchy can lead to new tables not having the correct number of clone descendant records, causing errors when the tables are queried.

    Refer to the listed KB article for details.

    Configuration Management Database (CMDB)

    PRB1787678

    The CSDM Life Cycle Migration 'bulkPopulate' method can be initiated multiple times, causing issues The 'bulkPopulate' method in the LifeCycleUtil script include can cause issues if it's directly called multiple times in a short time window. Copies are created of itself as they both try to update the same target CI records, and potentially overwriting changes made by each other.

    Content Experiences

    PRB1741977

    Script include and server errors are visible on the homepage Errors visible on Cloud Services's 'Browse Cloud Catalog' page.
    1. Log in as an admin user/role.
    2. Navigate to Cloud Services > Browse Cloud Catalog.

    Observe that script include and server errors are visible on the homepage.

    Content Experiences

    PRB1743499

    KB1641446

    With slow internet/fast multi-clicks, if a user multi-clicks Checkout or the Save and Publish button, it creates multiple draft and retired content There should be only a single draft version and retired versions should be unique.

    Refer to the listed KB article for details.

    Database Persistence - Data Access

    PRB1789910

    'Order by' columns are not added to the 'Group by' clause when joined on a denormalized table with aggregation via 'RLQuery' An SQLException error is thrown.

    Database Persistence - Data Management

    PRB1786870

    Remove the Toggle DMTableCleaner button
    1. Navigate to the list view of the sys_dm_job table.
    2. Navigate to the form view of the 'DMTableCleaner' job.

    Expected behavior: The Toggle DMTableCleaner button should be removed along with the script include it calls.

    Actual behavior: The Toggle DMTableCleaner button displays.

    Database Persistence - Data Scale

    PRB1715563

    Adjust connection opening/closing based on resource utilization and idle time Idle instances should only use 1-2 connections instead of 8-12. The connection closing doesn't really take effect currently.
    1. Converge an idle DBI and app node.
    2. Wait until it's truly idle.

    Observe that it still uses 10 connections, because the scheduler workers keep 8+ connections active all the time.

    Database Persistence - Data Scale

    PRB1788006

    DBLazy writer accumulates messages in an unlimited fashion in a thread stack In DBLazyWriter.process, it accumulates messages dequeued for triggering listeners later. If the queue isn't empty, it continues the loop forever and keeps adding messages to the arraylist. This is unbounded and runs into an out of memory error.
    1. Write to DBLazyWriter at a high rate to keep the queue full.
    2. Ensure that the DBLazywrite process runs slower so the queue is always full.

    Observe the memory usage of the Glide instance.

    Database Persistence

    PRB1685412

    Ordering by the Function field on a union query generates a query with syntax errors
    1. Create a table rotation/extension or use an existing one (such as sys_event).
    2. Create a Function field on that rotation/extension.
    3. Navigate to the 'List' view for the base table which would union all rotations (/sys_event_list.do).
    4. Add a sort on the Function field created in step 2 and add it to the list view if it isn't there.

    Notice the syntax errors.

    Database Persistence

    PRB1714908

    GlideAggregate that queries against table rotations generates incorrect results The addition of 'count_of_102865796' in the GROUP BY causes the different counts from the various table shards to be distinct values, instead of summed together. Removal of this field from the GROUP BY returns expected results.

    Discovery

    PRB1712133

    KB1648259

    There's a Large Table Handling (LTH) warning in queryPrivAffinity in the 'Insert Discovery Affinity & Cred Aliases' business rule when cmdb.name is null The encoded_query from the queryPrivAffinity function of the 'Insert Discovery Affinity & Cred Aliases' business rule lacks a control mechanism, causing a large table handling when the hostname is undefined.

    Refer to the listed KB article for details.

    Discovery

    PRB1767189

    The use_optimized_counter flag in DiscoveryStatus can cause started/completed counts to be incorrect and the status is set to 'complete' before its done The problem is difficult to reproduce on demand but it happened frequently on datacenter. It was noticeable there when they discovered a single Linux server. A sensor launched 22 downstream probes and deferred updating the start count until all probes had been launched. Sometimes, one of the probes would complete and its sensor would complete processing at the same time, launching all 22 probes as completed. The thread that was launching probes would increment started by 22 and then thread that processed the sensor would overwrite the started count and unto that increment.

    Discovery

    PRB1776300

    System event processing is held up by numerous running.process.absent events from Discovery and Agent Client Collector Visibility (ACC-V) In Tokyo, a new type of event was added to clean up orphaned application CIs and relationships after the table cleaner had deleted the cmdb_running_process records. The event is triggered by the 'RunningProcessReconciler' script include, as part of Discovery sensors, and Agent Client Collector Visibility checks. It is then processed by the tcmdb_ci_appl CIs as a 'Absent' status, which potentially involves looping through a lot of records. Where there is a big ACC-V implementation, this is triggered for each agent install every 8 hours by default, which leads to many of these events. These events use the general event queue and the 'Events Process 0' job, which leads to backlogged event processing where all other events have to wait for these events to clear before running.

    Document Management

    PRB1790213

    Add telemetry changes for PDF generation Add the following telemetry changes for PDF generation: export functionality and Html2PDF conversion, including number of pages, the output PDF file size, and generation time.

    Edge Encryption

    PRB1773542

    A scheduled upgrade fails for a proxy running on Windows It appears that the upgrade mostly completes. There's no errors in any of the logs, but the new Edge proxy service isn't started.

    Email Notifications

    PRB1681360

    A scheduled report isn't including attachments on all emails Email attachments aren't sent in emails that are split due to multiple recipients. The problem only occurs when the attached file size is larger than 220 KB.
    1. Create a scheduled email of reports that generates an .xlsx attachment over 180 KB.
    2. Add 2 recipients to the list.
    3. Set the glide.email.smtp.max_recipients property's value to 1.
    4. Trigger the report on demand.

    The attachment is only present in one of the split emails.

    Email Notifications

    PRB1763775

    KB1699112

    When composing an email in Service Operations Workspace (SOW) and copy/pasting the 'Caller' name to the To or CC field, the auto-complete for the reference field isn't appearing For example, when the user copies and pastes the name 'Abraham Lincoln', it displays the error 'Local address contains control or whitespace. Characters in format [a-z][A-Z][0-9] {-} are accepted'.

    Refer to the listed KB article for details.

    Email Notifications

    PRB1783999

    The 'Instance Email Reader' job stalls when only deleting (not retrieving) messages from POP3 When an instance is collecting mail, it retrieves messages at a significant rate under normal operations. When choosing only to delete messages (and not retrieve them), each mail reader job processes only 20 messages and then exits, forcing the instance to wait for the next scheduled email reader job 2 minutes later. This cycle repeats until the instance finds messages it chooses to retrieve, instead of just delete.

    Employee Document Management

    PRB1764655

    The Move attachment button isn't loading topic details in UI16 When using the Move attachment button from the Employee Documents plugin in UI16, the topic aren't loading. The issue also occurs on Next Experience.
    1. Provision an instance with the Employee Document plugin installed.
    2. Add some attachment to an HR case sn_hr_core_case.

      See that the Move Attachment button appears.

    Observe that when the button is selected and a user tries to change the topic detail, it doesn't load.

    Event Management

    PRB1760429

    A race condition in tag based clustering engine when multi job/node event processing is enabled is causing multiple tag clusters to be created This causes multiple tag cluster groups to be created for the same alerts.
    1. Set 'not for correlation' from the query job (set the property as false).
    2. Enable multi-job (4)/node event processing.
    3. Create a TBAC definition.
    4. Send several events at the same time with different message keys.

    Observe that multiple tag cluster groups containing the same alerts are created.

    Event Management

    PRB1772901

    There's an out of memory error in Alert Priority when an alert has high number of child alerts In the warning logs, there's a message: 'Size of Alert group has reached limit of..."
    1. Create 3 million events that would generate 3 million alerts.
    2. Configure the alert correlation such that 30k alerts would have a single parent and around 100 parent alerts each have 30k child alerts.
    3. Ensure that the parent alerts are in 'em_alert_trigger_queue'.
    4. Run the 'Event Management - Alert Priority Queue' job.

    Event Management

    PRB1784562

    Setting the main alert for Virtual Alert isn't consistent with the ACR script The ACR script chooses the main alert to be the first one (by initial_remote_time) with the lowest number but ignores severity. Setting the main alert for Virtual Alert in VirtualAlertHandler.java is based on severity, but the alert number is ignored. This is causing redundant Virtual Alert groups to open.
    1. Enable the TBAC property for debugging: 'evt_mgmt.sn_em_ tbac.log_acr_script_debugging'
    2. Send multiple events at once with the severity set to 'Critical', that eventually should be grouped together in a tag cluster.
    3. Verify in the 'Activity' notes of the created Virtual Alert that the main alert is consistent and doesn't change.

    Event Management

    PRB1786275

    On the 'Link' view, getHopsBetweenCIs fails on CMDB grouping Users are also receiving an error.

    Fixed List Queries

    PRB1721024

    Users are unable to create a record on a fixed query list The issue comes from a null string.
    1. Load a kb_knowledge list with a fixed query.
    2. Create a record using the New UI action.

    Expected behavior: A 'New record' page should appear.

    Actual behavior: An empty page appears.

    Flow Engine

    PRB1781775

    The FlowCallbackExecutor callback throws a null pointer exception FlowCallbackExecutor. executeCallBack throws a null pointer exception. It could be due to the usage of collectors.toMap, which breaks when there's null output values. This blocks major flows.

    Form Controller

    PRB1761770

    There's main and Washington mainline Oracle zBoot errors Error details on the main Oracle zBoot: 'StatementBatcher SEVERE *** ERROR *** Exception during batch statement commit to glide...'

    Form Controller

    PRB1761801

    A page is unresponsive when opening a list that is part of an activity stream in a legacy workspace sn-form handles list actions in ColumnLayoutElement that are dispatched by a component that the form doesn't own.
    1. Log in to a Washington DC instance.
    2. Activate the sys_aw_master_config record for Agent Workspace.
    3. Navigate to /cache.do.
    4. Navigate to Agent Workspace.
    5. Open an incident with an attachment.
    6. In the activity stream, open a list via the three-dots.
    7. Select the three-dots again to hide the list.

    Observe that the page becomes unresponsive.

    Hermes (Family)

    PRB1786830

    Updating the partition count in a topic causes the associated 'SUBSCRIPTIONS' to fail
    1. Create a topic and keep the number of partitions as 1.
    2. Create a consumer.
    3. Create a kafka stream with the topic created in the previous step.

      A subscription is created.

    4. Produce messages to the topic.
    5. Confirm the msgs are consumed by the consumer.
    6. Update the number of partitions from 1 to 4.
    7. Save the topic record.

    This changed value appears well in the topic page. Navigate to the 'Topic Inspector' page, and there's the increased number of partitions. But, in sys_kafka_subscription, the number of partitions still displays 1. Also, the subscription that is created in step 3 gets deleted and a new one gets created, every minute or so. The column 'Has Error' value changes from 'false' to 'true' and vice versa.

    Horizon Component Library

    PRB1733120

    The 'Customer activity' page loads slowly in a workspace Users are seeing this issue Intermittently on the first load and when loaded multiple times. Automated tests are failing due to this reason.
    1. Create a case with the Contact/consumer field.
    2. Navigate to the 'Customer activity' page.

    Expected behavior: The 'Customer activity' page should load with an activity feed and facets.

    Actual behavior: The page isn't loading with the activity feed and it's stuck in a loading state.

    HR Service Delivery

    PRB1772519

    After upgrading to Vancouver, COE Security has a different behavior For a set of cases, the user defined no read policies but only a write policy, because they want to restrict only write but not read. With this setup, all agents can read those cases but only some (as defined in policy) can write. After the upgrade, only the agents with write are able to read. The other agents who were able to read the cases before lost the access post-upgrade.
    1. Provide a user with write access via the COE security policy.
    2. Post-Vancouver, verify if no read policy is made that only the agents with write access are able to read.

    Incident Communications Management

    PRB1751178

    KB1638819

    There's errors in logs for using instanceof for ConferenceParticipantsTCM0 There shouldn't be any errors in the console and the isConferenceLeader method should return the value based on the input. However, the following errors are logged in syslog: 'Error in ConferenceParticipantsTCM0.getGroupMembers: JavaException: com.glide.script.fencing. MethodNotAllowedException: Function instanceOf is not allowed in scope sn_comm_management: no thrown error'.
    1. Provision an instance with the TCM plugin installed.
    2. Set up 'Notify'.
    3. Initiate a conference call from Service Operations Workspace.

    Observe that there's errors in the log.

    Install Base Management

    PRB1786211

    KB1650430

    The 'Account Page (Case View) Sold Product' related list must be updated with the default filter condition The 'Sold Product' related list displays as empty when upgrading from Vancouver to Washington due to additional filters. The issue is observed only when the records are associated with a product offer that has a specification set.

    Refer to the listed KB article for details.

    Instance Clone (Family)

    PRB1774663

    The user observes the alert 'General Data Exception detected by Database (No value specified for parameter 1.)' after submitting a clone on the 'clone_instance' page This isn't impacting clone progress, but it gives false alarm.

    Submit a clone via the 'Legacy clone request' page.

    An error warning displays on the top of the page.

    Instance Data Replication (IDR)

    PRB1787215

    For conserving disk space, remove the functionality of the seeding producer job waiting for ACK messages from the 'consumer' to 'publish next batches of data' It adds an unnecessary point of failure of, if the ACK messages aren't received or sent without errors, then jobs hang. The only benefit of the ACK message exchange was controlling the disk size. Disk size unchecked growth can be controlled at the kafka level with data retention policy settings.
    1. In QA env, stop the metadata consumer job on producer.
    2. Request new seeding on the consumer.
    3. Reset the metadata topic offset to end on the consumer.

    Notice the 'Seeding produce' job hangs.

    Integration Hub

    PRB1668651

    KB1645922

    Retry policies for connection timeouts don't work over HTTPS If users try to configure a retry policy for a connection timeout, it doesn't trigger when the request is done over HTTPS. The following exception is thrown in the UI: 'java.net.SocketTimeoutException: connect timed out'.

    Refer to the listed KB article for details.

    Integration Hub

    PRB1780194

    A data stream action fails with 'Undefined Schema' when a dynamic object is used as an output in a Washington release The issue is happening only on the Washington release where the parent plan runs on engine v2 and the child plans run on engine v1. The issue is not reproducible when both the child plans are run on engine v1 (Washington) or when both the plans are run on engine v2 (Xanadu).
    1. Open a Washington instance.
    2. Create a simple data stream action requesting incident records.
    3. Set the output as a dynamic object.
    4. Create a 'Get ServiceNow Object Schema (Dynamic)' action.
    5. Run/test the data stream action from a flow/subflow.

    Observe that the action/flow doesn't pass successfully.

    Integration Hub

    PRB1782980

    The script-side StrictReject map for the glide_hub_flow_ engine_glideencrypter crypto module causes an 'access denied' error Strict reject should be turned off for password2 or encryption support.
    1. Create an action with a script step.
    2. Decrypt the password field using GlideEncrypter in the script step.

    Expected behavior: The flow with the script step should be executed by creating a map for 'global.com_glide_hub _flow_engine_glideencrypter'.

    Actual behavior: Access is denied to the cryptographic module with a 'global.com_glide_hub_ flow_engine_glideencrypter' error on the flow execution.

    Inter-Instance Data Comparison Tool

    PRB1772226

    Remove the reseeding records limitation The current design retrieves the sys_id list that are either missing or mismatched records, or both. It creates partial seeding configuration with all sys_ids in the IN query as a string. It initiates the partial seeding. Since all sys_ids are added to a single string, the max sys_ids to be processed is limited to 100k sys_ids. In order to remove the limitation, improvements are being considered.
    1. Run DCT.
    2. Perform reseeding of mismatched records.

    Key Management Framework (KMF)

    PRB1762398

    The database view 'sn_kmf_crypto_user_policy' fails on Postgres instances The database view 'sn_kmf_crypto_user_policy', which joins sys_user_has_role and sys_kmf_crypto_caller_policy, fails on Postgres instances. Users can't correct the view manually, as the view is protected.
    1. On a Postgres instance, navigate to sn_kmf_crypto_user_policy.list.
    2. Verify that no records are returned.

    Language and Translations

    PRB1784816

    The translation loader does not dismiss store apps translations for deactivated languages Install a store app that carries translations in JSON format. Once installed translations for all languages are loaded.

    License Usage

    PRB1725703

    KB1648266

    The software model page does not open in Software Asset Management Pro after upgrading to the Vancouver release After upgrading to the Vancouver release, the software model page does not open when selecting the 'i' icon in Software Asset Management (SAM) Pro.
    1. Navigate to Workspaces then Software Asset Workspace.
    2. Select a record with a reference field and value.
    3. Select the 'i' icon in the reference field.

    Notice that the referenced record does not open.

    License Usage

    PRB1764672

    KB1644243

    License usage is slow for unlicensed entitles and progress indicator widgets are slow On the 'License Usage' page, there's widgets to show counts for ignored installs, unlicensed installs, and installs requiring action on the publisher result, product result, and software model result. These aggregated counts aren't stored and instead queries are executed, which can take a long time.

    Refer to the listed KB article for details.

    List Administration

    PRB1709691

    KB1641555

    Each sort order of a column on 'Dashboards' acts like a separate view In 'Dashboards', when a column is sorted, it seems like a new view is created with each column and its position in the list. The expected behavior is that the columns selected and their position remain the same regardless of what column is sorted.

    Refer to the listed KB article for details.

    List Administration

    PRB1789036

    Uptake the enableFocus property for internal uses of now-alert/now-alert-list within the 'Lists' component Include the enableFocus property to the message payload for any dispatches to the UXF notification handler where focus can safely be moved to the alert. Users with motor impairment often use only a keyboard and rely on quick and efficient navigation. This behavior doesn't let the banner be readily dismissed and forces to the user into extra tab stops. Additionally, the banner is covering focusable content.
    1. Log in to any instance.
    2. Navigate to any list within any workspace.
    3. Select at least one record from the list (select the checkbox) to enable the Export button.
    4. Select Export.
    5. Select Export again on the modal.
    6. Upon completion of the download, verify that there's a notification banner at the top of the page.

    Expected behavior: The keyboard focus lands on the X or Close button of the notification banner.

    Actual behavior: The keyboard focus is behind the banner. The user must tab multiple times to bring the focus back to the banner to dismiss it.

    MetricBase

    PRB1728947

    Registering a large number of triggers with Clotho causes some triggers to not get registered Series references are appending to a request URL when syncing triggers between Glide and Clotho. If there are too many series registered at once (for example 20k), the header becomes too large. Series references should be included in the request body so things are scalable.

    MetricBase

    PRB1785413

    'Accumulate' calls result in double counting A typo in GlideClothoClient results in a known series being resent alongside an unknown series. This causes the value doubling for them during 'Accumulate' calls.

    Mobile Platform

    PRB1754843

    When offline, there's missing redirected screens in the 'Chart' section in iOS Sections where offline mode is enabled aren't working as expected. The screens and sections don't load.

    Next Experience Unified Navigation

    PRB1768410

    Service Operations Workspace (SOW) tabs are less descriptive in Washington, with the workspace name missing Browser tab titles are less descriptive and not compatible with screen readers since the Washington DC upgrade. This happens for all workspaces, not just SOW.
    1. Log on to an instance.
    2. Navigate to Workspace > Service Operations Workspace.
    3. Verify that it correctly displays 'Home | Service Operations Workspace | ServiceNow'.
    4. Navigate to a list.
    5. Open an incident.
    6. Press back on the 'home' tab.
    7. Verify that it now displays 'Home | ServiceNow'.

    Next Experience Unified Navigation

    PRB1779558

    A domain with children whose entire hierarchy is turned off still appears in the domain picker
    1. Install Domain Extensions Installer with demo data.
    2. Navigate to domain_list.do.
    3. Turn off the entire hierarchy of a domain that has a descendant.
    4. Open the domain picker in the navigation header.

    Expected behavior: MSP doesn't display since the entire hierarchy is turned off.

    Actual behavior: MSP does display.

    Next Experience Unified Navigation

    PRB1787356

    KB1652699

    Some pages don't load when the property 'glide.ui.next_experience. instance_tools_disabled' is set it to false and Next Experience is turned off Pages such as 'Plugins', 'NLU Workbench models', 'Service Operations Workspace', and 'CSM-Configurable Workspace' don't load when the property 'glide.ui.next_experience. instance_tools_disabled' is set it to false and Next Experience is turned off. Admin role users are unable to access these modules: Plugins (now/app-manager/home), NLU Workbench models, Service Operations Workspace, CSM-Configurable Workspace. ITIL role users are unable to access Service Operations Workspace and CSM-Configurable Workspace.

    Refer to the listed KB article for details.

    On-Call Scheduling

    PRB1729558

    The canRead check on a table throws an error from a non-global scope For the tables 'on_call_user_contact_preference' and 'on_call_user_preference', if users check canRead access on the table as an admin from a scope other than global (eg. sn_cmdb_ws), the canRead returns a value but also throws a background error.

    On-Call Scheduling

    PRB1747212

    The on-call workflow is intermittently canceled at the 'Timer' activity without any error There's an intermittent issue where on-call workflows, like 'On-Call Assign by Acknowledgement per Rota', are intermittently canceling when the workflow reaches the 'Timer' activity, which is causing the escalation process to prematurely stop.

    Performance Analytics Dashboards

    PRB1774493

    Saving a report as a PNG/JPEG on a dark theme makes the text appear light grey in the downloaded file While saving report as image (using save as PNG or JPEG options) on the dark theme, the text appearance of the downloaded image is incorrect (light grey text, with an invisible title). This is happening for all reports when using the dark mode. When using the light theme, the downloaded image is correct.
    1. Navigate to a Washington instance.
    2. Open the 'Incident Overview' dashboard.
    3. On any report, select the hamburger icon.
    4. Select Save as PNG.
    5. Check the downloaded file for the texts appearing in proper font colors.
    6. Change the theme to dark from the preferences.
    7. Download the same report again.
    8. Check the texts in the downloaded file if they are appearing in light grey.

    Expected behavior: Text colors in the downloaded file in both themes should be in the proper colors.

    Actual behavior: Text colors in the downloaded file in dark themes are becoming light grey.

    Performance Analytics

    PRB1781018

    A legacy widget doesn't publish filters When the user has a custom filter in compatibility mode and tries to interact with it, the filter doesn't get applied.

    Platform Analytics Dashboard API

    PRB1781006

    Dashboards, Data Visualisations, and Scheduled Exports are set to have been created by the user who ran 'Platform Analytics Migration' instead of the original The 'created by' user was the user before migration who actually created the dashboards and scheduled exports. But post-migration, the 'created by' is changed to the person who initiated the migration.
    1. Navigate to a Washington DC instance.
    2. Have a non-admin user with roles like itil + report_scheduler, and impersonate this user.
    3. Create a report.
    4. Schedule the report and add it to the new dashboard.

      Entries in pa_dashboards, sys_report, sysauto_report should be created and updated by the user used in step 2.

    5. End the impersonation.
    6. Impersonate the user with a user with an admin role.
    7. Trigger 'Platform Analytics Migration', which should be completed without issues.
    8. Validate the new records created post migration in par_dashboard, par_visualization, par_export, par_export_visualization.

      The created by the user is set as the user who triggered the migration (the user in step 6).

    Expected behavior: Dashboards, Data Visualizations, and Scheduled Exports shouldn't have 'created by' set by a user who ran 'Platform Analytics Migration'. Instead, original creator names should persist.

    Actual behavior: Dashboards, Data Visualizations, and Scheduled Exports are set by the user who ran 'Platform Analytics Migration' instead of keeping the original user.

    Platform Analytics Migration API

    PRB1793305

    If individual dashboards have been rolled back, migration rollback fails an activation step
    1. On a Washington DC instance, migrate all dashboards.
    2. Rollback a couple dashboards from par_dashboard table.
    3. Set the rollback property to true.
    4. Perform a bulk rollback.

    Expected behavior: Bulk rollback actives everything properly.

    Actual behavior: The rollback fails.

    Predictive Intelligence Clustering Solution

    PRB1769445

    An error is thrown in the syslog by the job 'Application Suggestion - ITOM discovery' An 'Illegal parameter: null:' error is thrown in the syslog by the job 'Application Suggestion - ITOM discovery'. There's no impacts to processing. It's a redundant error message thrown 4 times every time the job runs.
    1. Run the scheduled job 'Application Suggestion - ITOM discovery'.
    2. Check the system log for errors.

    Predictive Intelligence for Knowledge Management

    PRB1783818

    The 'Related Article' widget isn't displaying knowledge articles from the 'Knowledge Similar Articles' solution Related articles aren't displaying due to ML prediction failing. This happens due to a validation error, as not all input columns defined in the solution definition are passed.

    Process Mining

    PRB1775603

    Licensing API blocks process mining for open instances for valid users

    Try using process mining on an open instance.

    Process Mining Workspace

    PRB1740430

    Users are unable to validate an automation request feature due to an error on the 'Automation request' tab The error: 'Please install Automation center plugin'.
    1. Navigate to the Process Mining application.
    2. Select All Projects.
    3. Open a project in the 'Draft' state.
    4. Select View in Workspace.
    5. Navigate to Analyst Workbench.
    6. Select Improvement Initiatives from the right side menu.
    7. Select the 'Automation request' tab.

    Expected behavior: Automation request content details should be available to the user in the automation request.

    Actual behavior: Users see an error 'Please install Automation center plugin'.

    Process Mining Workspace

    PRB1762674

    In Process Mining's 'Map Dashboard' widget, the incorrect theme can cause the process map to disappear
    1. Open platform-analytics-workspace > dashboards.
    2. Create a dashboard.
    3. Add a Process Mining map element.
    4. Select Edit on the 'Theming' section.
    5. Give a different name for 'Pill'.
    6. Save.
    7. Save and exit the edit mode.

    Observe that the process map disappears.

    Process Mining Workspace

    PRB1786987

    Set Objectives is slow to load and sometimes doesn't load at all
    1. From the left navigation, select Process mining > Projects > Create new.

      Observe that set objectives doesn't load and intermittently takes around 1 minute to load.

    2. Navigate to the 'Process Projects' page.
    3. Select Create new project.

    Observe that it doesn't load for the first time. If users close the tab and reopen by selecting the Create new project button again, it opens instantly.

    Record Watcher

    PRB1788115

    In rare cases, ResponderRegistryCleanerThread can remove all persisted responders from the memory when the database returns null on an exception RW Cleaner Thread wakes up every 10 seconds and does some cleanup tasks. As part of this, it tries to remove all orphaned persisted responders from memory. It gets all persisted responders from the database and compares them with what's in the memory. It removes those which are in memory but not in DB. In this case, because of an exception from the DB query, the database layer has returned null. The cleaner code understood this as there are no responders in database and removed all of them from memory, which has caused RW to not call the responder on a matching condition.
    1. Set glide.record_watcher.debug.extra to true.
    2. Ensure that 'sys_rw_action' contains a few entries. If not, create them.
    3. Shorten the interval glide.record_watcher.cleaner. orphaned_persisted_responders. cleanup_interval to 10 (seconds) for testing (so the breakpoint below 1 can be hit more often).
    4. Put a debug pointer at getPersistedRespondersFromDatabase before the actual query.
    5. When breakpoint hits, shut down the DB.
    6. Resume the program.
    7. Check the logs.

    Expected behavior: Responders are kept in the memory.

    Actual behavior: A log message like 'removing persisted watchers for tables: [list of all tables in memory]'.

    Schedule Optimization

    PRB1792462

    When an assigned task is canceled, it's returned as part of locked_tasks A locked task query was getting both canceled and closed tasks. Upon investigation, only closed tasks should be sent as locked.
    1. Create a work order task in the 'Pending Dispatch' state with the window_end date as the future.
    2. Perform graph QL query for the batch.

      The task should be returned as part of 'tasks'.

    3. Assign it an agent either manually or run SO.
    4. Cancel the same task.
    5. Perform a graph QL query for the batch.

    Observe that the task is returned as part of 'locked_tasks', which isn't expected.

    Script Includes

    PRB1761513

    Hung semaphore thread on EmbeddedSlotMap.java In Washington DC, there's been a large increase in the number of alerts for hung semaphore threads.

    Search Administration

    PRB1789281

    There's cache issues across domains when performing a global search There's cache issues across domains when performing a global search if the 'text_search' view on a list contains a short description in one domain and doesn't contain it in another domain.
    1. Navigate to any instance in a Washington DC version.
    2. Ensure that in the global domain there's a text_search view on the incident table.
    3. Set up a text_search view on the incident table in a domain other than global.

      Ensure to not include any string fields in the list layout.

    4. In this new domain, execute a global search against an incident that has previously not been searched on.
    5. Validate if 'No Title' or the sys_ID is displayed in the search preview.
    6. Switch to global domain.
    7. Complete a search for the same incident.
    8. Validate if 'No Title' or the sys_ID is still displayed.
    9. Try vice versa across both the domains.
    10. Verify that the expected behavior is obtained after clearing the cache (after cache.do).

    Server-side scripts

    PRB1730509

    In Firefox, after running a background script, the browser's Back button doesn't return to the script page After running any script on the 'Background script' page, when a user navigates back with the arrow in the Mozilla browser, they're redirected to the instance homepage and don't stay on the background script page.
    1. Navigate to the filter navigator, System Definition > Background Script.
    2. Type 'gs.log();'.
    3. Change the scope to global.
    4. Run the script.
    5. Select the back arrow from the browser.

    Expected behavior: The user should stay in the 'Background Script' menu.

    Actual behavior: The user is navigated to the instance home page.

    Service Level Management

    PRB1734205

    'SLA Repair' and 'SLA Timeline' using HistoryWalker give undesired results when audit data has a split transaction For example, the same update number for different record internal checkpoints.

    Service Mapping

    PRB1745336

    The property that limits the number of lookup items added to the service model should include cmdb_ci_network_adapter 'sa.lookup_tables. tables_for_limited_query' is a comma-separated list of table names to limit the lookup query. A limit is added to the table list, and all tables are inherited from them. The default value is cmdb_ci, cmdb_serial_number. 'sa.lookup_tables. exclusion_tables_f or_limited_query' is intended to complement 'sa.lookup_tables. tables_for_limited_query'. It's a comma-separated list of table names to unlimited lookup query. The table list is unlimited, and all tables are inherited from them. The default value is empty. 'sa.lookup_tables. max_records_to_check' involves the numbers of records to limit the lookup query. The default value is 100.
    1. Create many records (7-9 K) for a network adapter on any switch in cmdb_ci_network_adapter.
    2. Add this switch to an application server.
    3. Run recomputation of the service with this switch.

    An error displays: 'WARNING *** WARNING *** Large Table: Table handling an extremely large result set: 45321...'

    Service Mapping

    PRB1791752

    Put a hard limit on properties that can be destructive These are the properties: sa.max_relations_in_service, sa.service.max _ci_service_population, svc_by_tags.max_ci_limit. A business rule should be created that doesn't allow increasing the property values above twice the default value. The business rule name is: 'Limit' + the name of the property.
    1. Try to set the following properties:
      • sa.max_relations_in_service > 10000 (default 5000)
      • sa.service.max_ci_service_population > 2000 (default 1000)
      • svc_by_tags.max_ci_limit > 2000 (default 1000)
    2. Verify if these properties can be set.

    These shouldn't be possible to set as they can be destructive.

    Service Portal

    PRB1725506

    Attachments aren't displayed in Service Portal's activity stream Attachments that are added more than 8 hours prior to the submission of the record aren't visible in the activity stream on Service Portal. This restriction impacts the scenarios where the users use the 'Add to cart' or 'Add to Wishlist' functionalities in portal. There could be scenarios where the users add the attachment to the catalog item and then save the catalog item in the cart for future submissions.
    1. Log in to Service Portal.
    2. Add an attachment to a catalog item.
    3. Add the catalog item to the cart.
    4. Log out.
    5. Wait for more than 8 hours.
    6. Re-log in to the instance.
    7. Submit the catalog item.

    Actual behavior: The image/attachment added is missing in the activity stream on portal.

    Expected behavior: The attachment should appear in the activity stream.

    Service Portal

    PRB1745184

    'Service catalog' in the breadcrumb isn't translated consistently The issue is reproducible in Vancouver in the 'Breadcrumb' widget. Other texts are translated.

    Service Portfolio Management

    PRB1750190

    The document ID doesn't handle the table extension transparently If a record is moved from a 'Base' table to an 'Extended' table (while keeping the sys_id), the expectation is that existing references via the document_id (e. g. sys_attachment) or related features (e. g. sysapproval_approver) aren't broken. However, the actual behaviour is that references via document_id are not brought forward to the extended table.
    1. Create a record in cmdb_ci_service.
    2. Attach a file to it.
    3. Select the related link 'Convert to Business Service'.
    4. Check the attachment.

    Expected behavior: The attachment is still there.

    Actual behavior: The attachment isn't there.

    Software Asset Reconciliation

    PRB1734615

    SuiteEngine performance improvement The user has 1.5m Microsoft installs that were taking 3 hours to complete.

    Software Entitlements

    PRB1728056

    Microsoft License Statement (MLS) bulk import creates duplicate entitlements Errors are created.

    Software Lifecycles

    PRB1756991

    A life cycle report should be able to handle empty life cycle dates

    System Export Sets

    PRB1792544

    KB1699358

    If Log Export Service (LES) is enabled and the LES source is configured with syslog in a global scope filter, there's a null pointer exception when using gs.log after a Washington DC upgrade There's an error: 'Script execution error: Script Identifier: null.null.script, Error Description: java.lang.NullPointerException, Script ES Level: 0 Couldn't decipher the stack trace resulting from the following JavaScriptException...'

    Refer to the listed KB article for details.

    Time Card Management

    PRB1785263

    Managers are unable to copy timecards When a manager attempts to copy time cards for one of their reports, the system returns {0} time cards created.
    1. Impersonate a manager in an instance.
    2. Select a user from the user list.
    3. Select the Copy time card button.
    4. Select the date for the time sheet.
    5. Select Okay.

    Observe the error message.

    UI Field Administration

    PRB1695552

    For translated text reference fields, the reference pop-up list doesn't display a value when a right-to-left (RTL) language is used
    1. Have any RTL language, like Arabic, plugin installed.
    2. Create a table to be used as a reference table with at least the following fields:
      • Name: Display column, Type: Translated Text, Display: true
      • Name: Active, Type: True/False
      • Name: Type, Type: Choice, Choices: Type A, Type B
    3. Create at least the following records in the new table:
      • Display column: 'Ref - Type A - Active 1', Type: Type A, Active: true
      • Display column: 'Ref - Type A - Inactive 1', Type: Type A, Active: false
    4. Create a reference type of field on the incident table.
    5. Make that field refer to the newly created table in step 1.
    6. Configure the incident table form layout to show this field.
    7. Change the session language to Arabic.
    8. Open any incident.

    Expected behavior: The type 'Ref' in the Reference field should display the 2 entries from the table created before.

    Actual behavior: The type 'Ref' in the Reference field displays the 'Reference' list but the list has empty entries.

    UI Form Administration

    PRB1773340

    A field hint no longer displays for a boolean field when a field label URL is configured This works as expected in a Vancouver instance. The issue is only with checkboxes and read-only choice fields with URLs configured.
    1. Log in to the latest Washington DC release.
    2. Enable User settings > Display > Show help tips on forms.
    3. Open a problem record.
    4. Add a Knowledge field to the form layout.

      Notice the hint is added to the field label when hovering over the [?] icon.

    5. Right click on Knowledge field > Configure label.
    6. Add a URL.
    7. Update.
    8. Hover again over the [?] help icon for the field.

      Notice the hint isn't displayed.

    Expected behavior: Both the URL and field hint should work.

    Actual behavior: When the URL is added, the field hint breaks.

    UI Form Administration

    PRB1783850

    KB1651462

    High memory usage by viewrule_all_cache (Workspace View Rules) The viewrule_all_cache can consume excessive memory on application nodes. This can cause memory contention/increased garbage collection/application node performance degradation.

    Refer to the listed KB article for details.

    Upgrade Center

    PRB1776398

    After a family upgrade and activating the Store app, conditional content isn't being installed
    1. Provision an instance in Washington DC.
    2. Upgrade the instance from Washington DC to Xanadu.
    3. Upgrade the Security Center (SC) Store app from 1.5.3 - 1.6

    Expected behavior: The 'Client push notifications' table exists and the glide.cmdb.canonical.url property is displayed.

    Actual behavior: The sn_vsc_client_push_notifications table, and others, are missing.

    Usage Analytics

    PRB1771649

    Enforce basic tracking for users who consented to the decision to 'NO' and add a privacy link on the 'Enable/Disable' page
    1. Log in to a Glide instance with a valid user.
    2. Ensure that instance level usage analytics is on.
    3. Turn off the user consent policy for that user

    Actual behavior: Heartbeat isn't sending.

    Expected behavior: Heartbeat should send.

    UX Framework

    PRB1742312

    The test step to open a new form gives the error 'GraphQL API - GlideGraphQL : documents list does not contain valid document records: com.glide.graph' There's an issue with the 'Open a New Form' ATF test step. The first time the Activity Stream GraphQL API is called, it's with the sys_id=-1. Although this doesn't return a valid record, '-1' is understood to be a marker for a new record. As such, in this instance, the Activity Stream GraphQL API returns 'null'.

    UX Framework

    PRB1792196

    The 'Extend your session' option isn't occurring for single sign-on (SSO) users After upgrading an SSO-enabled instance to Washington DC, users no longer see the 'Extend your session' modal on the 'Platform' view. Instead, users are logged out after their session times out. This causes work to be lost. The modal still appears in the /esc and /sp portals, however. For the 'Extend session' modal to appear for SSO users, glide.ui.session.show _modals.override must be created if it doesn't exist and set to 'true' in the system properties 'All' menu.
    1. Navigate to an instance with read/write access via SSO.
    2. From the filter navigator, navigate to Content Library > View All.
    3. On the next screen, select New in the top-hand corner.
    4. In the modal that pops up, select Rich Content.
    5. Select Continue.
    6. In the left-hand column, in the Content name field, enter any value.
    7. Select Open Editor.
    8. Select and type in the card.

    Observe that after 15 minutes, the user is logged out automatically and no modal populates.

    UX-Metrics

    PRB1772426

    UX-Metrics must enable basic tracking for users who consented to the decision 'NO' A privacy link should be added to the 'Enable/Disable' page.
    1. Log in with a Glide instance with a valid user.
    2. Ensure that instance level usage analytics is on.
    3. Turn off a user consent policy for a specific user.

    Expected behavior: When turning off the user consent policy for a specific user, heartbeat should send to appsee backed.

    Actual behavior: Heartbeat isn't sending to appsee backed.

    Virtual Agent Designer

    PRB1793134

    A user without elevated privileges can't create a large language model (LLM) sys_cs_context_profile
    1. Impersonate a user that doesn't have elevated privileges, but has the virtual_agent_admin role.
    2. Navigate to /sys_cs_context_profile.do.

    Observe the model_type field is NLU/keyword and can't be set to LLM.

    Virtual Agent Designer

    PRB1793793

    NAVA topic execution doesn't populate the 'API' column in a sys_cs_fdih_invocation table, causing ETL job failure For regular Virtual Agent conversations, topic execution populates an 'API' column in a sys_cs_fdih_invocation table, which is used by the ETL job to process the conversations. However, for NAVA conversations, topic execution doesn't populate the 'API' column, which leads to ETL job failure, causing the Virtual Agent dashboard to not display all conversations.
    1. Set up NAVA.
    2. Search 'What is spam'.
    3. Check the sys_cs_fdih_invocation table for the 'API' column of the newly generated record.

    Expected behavior: The 'API' column is populated.

    Actual behavior: The 'API' column is empty.

    Virtual Agent

    PRB1770948

    In Virtual Agent, an action utility doesn't display all the relevant flow actions Flow Designer actions are missing an action utility in Virtual Agent Designer.

    Virtual Agent

    PRB1781835

    Unable to create a new Now Assist deployment channel record for the provider channel 'Identity'

    Virtual Agent

    PRB1782232

    Utterance length isn't checked at topic discovery, which allows large utterances that cause a slow down In topic discovery, no input length is checked, so large utterances can be passed into the search. Also, USER_INPUT_MAX_LENGTH isn't checked and is also too large for topic discovery.
    1. Provision an instance with the glide.cs.chatbot plugin.
    2. Start a conversation with a virtual agent.
    3. Enter a large utterance at topic discovery.

    Observe slowness.

    Virtual Agent

    PRB1789855

    In Virtual Agent after a Washington DC upgrade, every input that matches more than one keyword throws an error Virtual Agent should display all topics that match a keyword, but it displays a technical error message or occasionally hangs.

    Virtual Agent

    PRB1791956

    Multiple entities for slot filling cause a conversation to crash When an NLU-based user has multiple entities in a conversation being filled during NLU topic switching, a choice picker is presented. After a choice is made VA displays 'I'm having technical issues and won't be able to continue this conversation'.

    Virtual Agent Web Client

    PRB1768332

    After an upgrade, Now Assist in Virtual Agent results aren't formatted Newly provisioned instances have the same problem.

    Work Order Management

    PRB1774985

    Users are unable to unassign the tasks from a service organization (SO) solution for some tasks, which is causing double booking
    1. Run optimization for a territory (with batch).
    2. Add or update some tasks.
    3. Run optimization again.
    4. Verify that there are unassigned_task_ids in the solution file.

    Observe that the tasks were not unassigned due to the validation in FieldServiceTerritoryPlanning. validateAgentMembershipForTerritory.

    Fixes included