Yokohama Patch 13

Release version: Yokohama

Updated May 5, 2026

57 minutes to read

The Yokohama Patch 13 release contains important problem fixes.

Yokohama Patch 13 was released on May 5, 2026.

Build date: 04-29-2026_0346
Build tag: glide-yokohama-12-18-2024__patch13-04-21-2026

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

Yokohama Patch 13 includes 297 problem fixes in various categories. The chart below shows the top 10 problem categories included in this patch.

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

Changes in Yokohama Patch 13

Define which request headers are expected for each resource within a scripted REST API.

Associate scripted REST API query parameters with a resource.

Use sample requests made to an API resource to generate request header associations, query parameter associations, and a request schema for that resource.

Available system properties

glide.rest.sampling.enabled

Enables request sampling from scripted REST API resources. REquest samples can be used to automatically generate definitions for request headers, query parameters, and schemas. General guidance to help limit performance problems is to only enable this property in non-production instances.

Type: true | false
Default value: false
Location: System Property [sys_properties] table

Define one or more REST API response headers in a REST API.

Specify the expected data structure for requests or responses within a REST API by defining a schema.

Add a schema and relevant response headers to a REST response to define the response's expected data structure.

Add a schema to a REST request record to define the request's expected data structure.

Define available query parameters to control what values a requesting user can pass in the request URI.

Define scripted REST API request headers to control which headers the API accepts.

The guarded script evaluator enhances instance security by supporting only a restricted scripting language and detecting or rejecting untrusted scripts that use unsupported JavaScript features.

Review the JavaScript APIs that guarded script supports to help you analyze scripts in the Incompatible Guarded Scripts list and either rewrite them or create an exemption for them.

Now Assist and agentic AI release notes

For Now Assist new features and changes, see Now Assist and agentic AI release notes.

Review scripts that are incompatible with guarded script and either rewrite them to use supported features or create an exemption for scripts that can't be rewritten.

The script sandbox evaluator helps prevent executing untrusted scripts on an instance by limiting the APIs available to scripts.

Scripts that run in the script sandbox evaluator can use features supported by the JavaScript engine and the sandbox environment, except for certain restricted methods. Untrusted scripts are processed by the script sandbox evaluator under the following conditions:

A script has been granted a guarded-script exemption (manually or automatically).
When guarded script is in Phase 1: Detection, and a script is sent to the server by an authenticated user.

For more information about guarded-script exemptions and enforcement phases, see .

Schemas define a structure that can be used for API requests and responses, including data type, expected fields, and formats. You can define multiple schemas within a scripted REST API, which can be used to specify request and response contents of the resources within that API.

Schemas must be defined using OpenAPI version 3.0.1 format.

The script sandbox environment is a restricted execution context in which untrusted, client-generated scripts run on the server using one of two evaluators: the guarded script evaluator or the script sandbox evaluator.

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
Application Manager PRB1889269	Progress worker state isn't set to complete after worker is finished	Installing plugins shows an error even when successfully completed. In the logs, it looks like the error comes from a script include. Previously, the code block was a part of a larger function that had appVersion and appID as parameters. The code was refactored into the forceSyncWithStore function, which doesn't have appVersion and appID as parameters. If the gs.info line is reached, it throws an error because those two fields were not passed into the function.	Install the custom scoped app (e.g., x_conym_atf_sc_req, x_conym_hrp). Check 'sys_upgrade_history' to confirm the installation is marked as 'complete'. Check 'sys_progress_worker' list for the message 'Progress worker state is not set to complete after worker is finished.' Check 'syslog' for the error 'com.glide.script.RhinoEcmaError: 'appVersion' is not defined.'
Application Manager PRB1981044 KB2719949	The latest version in the sys_store_app table is updated to the installed version after an install, thereby causing the Upgrade button to be unavailable in App Manager	Two functions (_fixLatestVersionFor WithdrawnInstalledApps, _fixLatestVersionFor WithdrawnInstalledCustomizations) in Application Manager's UpdateChecker. checkAvailableUpdates API (aka Sync in the UI) were missing a required GlideRecord.addQuery constraint when the 'sn_appclient.enable_ app_manager_checksums_cache' sys_property was set to true (default). This issue led to the latest_version being set to the incorrect values on Install/Sync.	Refer to the listed KB article for details.
Database Persistence PRB1971863	Transaction canceled: 'Failed to get database connection due to exhaustion of connection pool on the node'	A spike in requests can produce a demand exceeding pool max capacity. Some connections are created, but the process is slow, causing various race conditions and delays and eventually causing a timeout in one operation.
Email Notifications PRB1952527 KB2743221	An email template isn't applied when users select Apply template in the mini/full composer	The template isn't applied.	Refer to the listed KB article for details.
MID Server PRB1969766 KB2664275	Missing MIDLogFileHandler MID Server properties cause MID servers to go into an infinite loop of restarting during start up	MID Servers repeatedly come up and then go back down for a few seconds before the MID Server comes back up again. This loop repeats continuously if certain MID Server Properties are missing from the ecc_agent_property table on the instance.	Refer to the listed KB article for details.
Multi-provider Single Sign-on (SSO) PRB1979071 KB2715489	SAML SP Keystore expiration		Refer to the listed KB article for details.
UI Form Administration PRB1998771 KB2898934	Installing the com.glide.ai_record_activity plugin doesn't create the sys_ai_record_activity table, causing NullPointerException errors	The user observes errors getting records for sys_ai_record_activity: java.lang.NullPointerException.	Install the GenAI Q4 feature or upgrade an existing instance. Ensure the plugin com.glide.ai_record_activity is installed. Observe the system logs or error output about an error getting a record for sys_ai_record_activity: java.lang.NullPointerException.

All other fixes


Problem	Short description	Description	Steps to reproduce
Activity Stream PRB1996382	When setting glide.ui.escape_text to 'false', forms break		Open any Zurich instance. Upload the sys_properties file for glide.ui.escape_text. Run a /cache.do. Load any relevant record, such as 'Incident'. Expected behavior: The record should load normally with the glide.ui.escape_text sys_property activated. Actual behavior: A blank screen with a basic message at the top calls out the ampersand issue.
Activity Stream PRB1998124	Activity stream citations cause an increased page load time		Navigate to an instance. Open Service Operations Workspace. Open any incident record with a large amount of activity stream entries (~2000 entries). Open the browser's dev tools window. Select the 'Performance' tab. Select the Record and reload button. Expected behavior: The page load time shouldn't be affected by CitationHighlightBehavior.js and require less time. Actual behavior: The page takes about 6 seconds or more to fully load, and the 'Performance' tab displays that it is due to CitationHighlightBehavior.js.
Activity Stream PRB2004286	A user's filter is incorrectly keying the supplemental map with translated keys, leading to an empty activity stream after changing the language	A supplemental map is incorrectly keyed with a translated string. It's also not correctly handling the case where the key in the map doesn't exist, leading to an uncaught exception and the activity stream not loading.
Activity Stream PRB2008384	The Copy button on journal entries doesn't always copy text or notify users correctly	This is reproducible in UI16 and Workspace.	Navigate to any relevant workspace. Open any incident record. Turn on the 'Rich text editor' in activity stream's Compose. Add a work note or comment to the activity stream with some of the text set to bold. Select the Copy to clipboard button on the new journal entry. Expected behavior: When rich text is copied, a notification should display saying 'Copied to clipboard'. When pasted into a plain text editor, there shouldn't be [code] tags around the text. Actual behavior: When rich text is copied, a notification doesn't display. When pasted into a plain text editor, there shouldn't be [code] tags around the text.
Advanced Work Assignment PRB1935541	Work item incorrectly moved to cancelled state instead of queued state	Without external routing, the rejected work item should move to the queued state. Instead, the work item moves to the cancelled state, and the interaction moves to closed_abandoned.	Change mweb sys_cs_channel to type 'Messaging'. Enable external routing for 'Chat - Asynchronous' Channel. Mark Chat Async Queue 'external' as true. Make two external agents available in workspace. As a system admin, start a chat in the SP/CSM portal. Once routing to live agent starts, a new work item is created. Assign the work item to the first agent using manual assignment API via REST API explorer. Change the first agent presence to offline. Send a message from admin. The conversation is reassigned, which moves the interaction state to New and creates a new work item. Assign the new work item to the second agent using manual assignment API via REST API explorer. Reject the work item. Expected behavior: Without external routing, the work item is moved to the queued state, looking for new agent. Actual behavior: The new work item moves to the cancelled state, and the interaction moves to closed_abandoned.
Advanced Work Assignment PRB1938992	High AWA assignment cycle times for impact voice call routing	A user is uptaking in NVC and they are facing slow response times for call routing. It's taking ~12 seconds to route the call to agent. Each AWA assignment cycle is taking ~11 secs and out of which ~8 secs is being spent for Agent Map creation. Expected response times should be less than a second as there's no other load in the system.
Advanced Work Assignment PRB1981019	AWA TransferToQueue shouldn't reassign a workItem to an AI Agent once rejected	If the interaction was previously transferred from the AI Agent, it shouldn't be transferred back to the AI Agent later (even if the AI agent is available).	Make sure an AI agent (sys_user with identity_type= 'ai_agent') is part of the agent chat assignment group of the chat queue. Make it AI available. Create a workItem. Assign it to the AI agent. Use the 'Transfer to queue' API to transfer the interaction to the same chat queue where the first workItem was created. Make the agent 'Beth Anglin' available to receive the transferred workItem. Invoke the Transfer To queue quick action from Beth. Select the same chat queue. Observe that the interaction is transferred to the AI Agent.
Advanced Work Assignment PRB1982818	After removing a service channel, the current universal capacity of the agent isn't decreased	After removing a service channel from the 'Available' presence state, the universal capacity for the existing documents related to that channel shouldn't be considered.	Create a service channel ('Chat-test'). Add it to the 'Available' presence state in awa_presence_state. Create related queues and assignment eligibility pools. Create a universal capacity record for the agent ('Beth Anglin'). Make the agent 'Available'. Create documents for that and let those be assigned to the agent. Observe that the current universal capacity of the agent is increased. Remove the service channel from the 'Available' presence state in awa_presence_state. Observe that the current universal capacity of the agent stays the same. In Xanadu and Zurich, the current universal capacity gets decreased.
Advanced Work Assignment PRB1985807	Graphql calls related to snOpenframeWrapper happen in regular spikes across all nodes and cause large CPU spikes	After the upgrade to Australia, a specific graphql call happens every 10 (or sometimes 5) minutes on HI.
AI Agents (Glide Family) PRB1975798	Glide fixes to JSON return type in AIAToolDataPillUtil		Ensure that 'capability' text is within the Value field apart of the sn_aia.supported_tools sys prop. Create or find a skill tool that takes JSON as input. Apply that skill to a tool and then add to an agent. Run the agent in the playground. Observe the execution log on the right hand side. Expected behavior: The user sees the actual JSON as an input for the skill tool. Actual behavior: The input is blank, causing the agent to fail.
AI Gateway - Security PRB1990717	AI Gateway Security Q1 Complete Feature Backport - Placeholder Story		This is a product update.
AI Search (Glide) PRB1934845	Make query boolean mode a first class parameter for search	When invoking AI search API, there should be a way to override the default query boolean mode. The goal is to first make NAVA default to OR mode.
AI Search (Glide) PRB1971388	IngestableDocument.getObjectSize isn't counting fEmbeddedDocuments	Indexing the document/table causes out of memory errors.
AI Search (Glide) PRB1972570	The display value in 'Location' gives the sysID rather than the country code	The issue occurs in the latest Zurich, Yokohama, and Now Assist instances. Sys_user.country is working but sys_user.location.country is not.
AI Search (Glide) PRB1990134	If all KBBs from one doc are removed from KB late binding, the post process stops and all following docs are returned	If ACL filtering removes all the embedded documents from a knowledge base article (and there's no embedded match), the code stops processing all subsequent documents.	Ensure the KB is using ACL instead of UC 'glide.knowman.search.apply_acls'=true. Create two KB articles: For KB1, the user has ACL access, but cannot access any of the KBB. For KB2, the user has no ACL access. As the user, search a query that matches KB1. (Or use result improvement rule to boost KB1.) Expected behavior: KB2 is not shown because the user has no ACL access. Actual behavior: KB2 is shown.
AI Search (Glide) PRB2006426	When a document with an attachment is removed by late binding, the next attachment where the user has access is also removed	When ds_document_version parent documents are filtered by an ACL and their children are skipped (the continue at line 516–518 bypasses the recursive child processing), the sys_attachment sparse cursor falls out of sync. It was built assuming every child would be processed in order, but skipped parents leave gaps.	Create 3 KBs with an attachment. Ensure that the KBs and attachments are relevant to a query. (KB1 is most relevant. KB3 is least relevant but still matches a query.) Add an ACL to block the second KB but allow the first and third. Force a late binding KB in ais_datasource or the 'everything by' property. Query in a search preview as a user. Expected behavior: KB3 should be displayed with an attachment. Actual behavior: KB3 is displayed but with no attachment.
AI Search (Glide) PRB2008247	A negative offset isn't handled properly, which causes no results after moving to the next page		Change the 'glide.ais.query.search_operator' property to 'OR'. Perform a search on Service Portal which returns more than 12 results. Select the Next Page arrow button. Observe that there's no matches.
AI Search PRB1823355 KB2301356	When there is no scope or global for a scoped admin scoped app, an invalidate query is sent	AI Search doesn't work as expected and returns no results when searched with any query. On search preview, the message 'An error occurred. Check the logs for more details and contact Support if this issues continues' pops up before the query can even be triggered.	Refer to the listed KB article for details.
AI Search PRB1838976	When Next Experience is off, searching for anything on the SOW/CSM workspace redirects the user to the homepage	This happens from Global Search. Instead of redirecting, it should display results using UI16.	Open a Washington DC or Xanadu instance. Turn off the Next Experience by setting glide.ui.polaris.experience to 'false' from sys_properties. Navigate to file navigator. Open Service Operations Workspace. Open the list from the left side panel. Select All under Incidents. Copy any incident number or search for any change/problem number. Search that number from the top-right search bar. Observe that it redirects to the homepage instead of displaying search results. If the user sets glide.ui.polaris.experience to 'true' from sys_properties, then it displays search results inside SOW in a new tab.
AI Search PRB1841579	Sn-search-combobox-desktop's use of createGraphQLEffect dispatches errorActionType due to a 401 error for public GraphQL endpoints	The batch request fails due to 401 even though the GraphQL request in the batch was actually successful.	Set the Service Portal home page to public. Set the Typeahead Search widget to public. Open Service Portal without logging in. Open Developer Tools. Select the Search Box. Observe in Seismic Dev Tools that the batch request failed due to 401. Then, observe in the Network tab that the GraphQL request in the batch was actually successful.
AI Search PRB1886370	AisHostnameProximityEstimator reads datacenter names from host names as case insensitive, but then compares them as case sensitive
AI Search PRB1896807	'Index All Table' from the list view of a datasource should trigger 'No block indexing' instead of 'Block reindexing'	This causes an event queue pile up when the user triggers multiple reindexing from the list view of the 'datasources' page.
AI Search UX PRB1943058	There's a global search highlighting issue when using a Safari browser in Next Experience with Zing	There's a visual rendering issue affecting the search bar on the Support ServiceNow portal when accessed via Safari OS version 26. The text typed into the search bar isn't properly visible regarding the theme, making it difficult for users to interact with the search functionality.	Log in to any instance. Select any available theme. Navigate to the support portal. Select the search bar and begin typing. Observe that the input text isn't rendered correctly, affecting visibility and usability.
AI Search UX PRB1980290	In the Service Portal with Now Assist enabled, when the tab focus is on the link and the 'Enter' key is pressed, the link doesn't open		Navigate to Conversational Interfaces > Assistants. Under Now Assist in Virtual Agent, select Display Experiences > Add service Portal > Save. Open any instance. Navigate to the Service Portal. Navigate to the search's Edit field. Type 'email' and search for it. Navigate to the search results > select a filter. Navigate to Any links in the 'Results' section. Attempt to press the 'Enter' key to select. Observe that when the tab focus is on the link and the 'Enter' key is pressed, the link does not open.
AI Search UX PRB1981121	Global search isn't working with zoom in/zoom out	Global Search in the Native UI doesn't return results when the filter navigator is pinned and the browser zoom is set to 125% or 150%.	Log in to an instance. Use the Native UI. Pin the filter navigator (left navigation panel). Set the browser zoom to 125% or 150%. Perform a global search for a known existing record. Observe that the search returns 'No Record Found'. Change the zoom to 80% and repeat the same search. Verify that the results appear normally.
AI Search UX PRB2004210	A Zing component isn't loaded with an error on the console: '__dyImp0(...).then is not a function'	Dynamic import fails as the component is already loaded by another module, in which case the dynamic import won't return a promise.	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.
AI Search UX PRB2006200	Synthesized answer loading card status text isn't fully visible in a narrow viewport when non-English locale text exceeds container bounds	When a user performs an AI Search query in a narrow viewport (reproducible via Chrome Responsive Design View on desktop) on Service Portal without enhanced chat, and with the session language set to a non-English locale, the synthesized answer loading card clips overlap its content. The loading status text and any action elements within the loading card might overlap together rather than wrapping or expanding the container height. The synthesized answer and the underlying search results continue to function correctly; the issue is limited to the visual presentation of the loading card during the analysis phase. The behavior is triggered by the longer text strings used for loading status messages in non-English locales. For example, the German loading status text 'Ihre Anforderung wird analysiert' (Analyzing your request) is significantly longer than the English equivalent, and the loading card container does not wrap to accommodate it in a narrow viewport.	Open an instance with ESC portal and Now Assist in AI Search with the session language set to German. Open Chrome and navigate to the AI Search page of ESC portal. Open Chrome Developer Tools and turn on 'Responsive Design' view (toggle device toolbar). Set the viewport to a narrow mobile width. Confirm that the instance language is set to German. Enter the search keyword 'übersicht fristen' in the search box. Submit. Observe the synthesized answer loading card as it renders during the analysis phase. 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.
Analytics Data API PRB1844011	Using 'trend by week' leads to incorrect record count when the new year starts
Analytics Data API PRB1928857	The maximum number of groups on the heat map visualization in Platform Analytics resets to 10 when filtering	After saving and refreshing, the column count resets to 10. Instead, all selected groups should remain visible.	Open an instance. Create a heatmap data visualization. Add a 'Group By' under Columns with more than 10 columns. Select Apply. Observe that all columns display correctly at this point. Save and refresh the visualization or browser. Expected behavior: All selected groups remain visible after saving or refreshing. Actual behavior: The column count resets to 10 every time.
Analytics Data API PRB1931476	Issue with Incident SLA donut translation in Service Operations Workspace (SOW)	After changing the language in SOW, the Incident SLA widget appears as 'Breached'. After clearing the cache, it then appears as 'undefined'.	Impersonate a user. Open SOW. Notice that the Incident SLA widget label appears as 'Breached'. Change the language to any language. Notice that the label still appears as 'Breached'. Clear the cache by cache.do. Notice that the label now appears as 'undefined' while the translation is already added in the sys_ui_message table.
Analytics Data API PRB1952604	The data label indicators show the same unit in Platform Analytics Data Visualization	After creating a data visualization that uses two indicators as data sources with different units, the data labels reflect the same unit.	Create a data visualization (line). Add to the data sources with different units (# and % for example). Show the data label. Expected behavior: The data label unit in each indicator reflects the unit of the indicator. Actual behavior: The data label units in both indicators are the same, and not respecting the configured unit of the indicators used as a base.
Analytics Data API PRB2001019	Accept 'group by' selection for all variable types	When comparing to classic reporting, many variable types are not available for selection in the 'group by' component in the Platform Analytics visualizations.	Navigate to Platform Analytics Experience. Create a dashboard. Add a donut visualization: Datasource: sc_req_item Group by: variables Search for: Sample item Notice only multi-choice and 'Select box' are available. Create the same visualization in classic reporting. Notice that there's a lot more options available.
Analytics Export API PRB1896576 KB2280342	Inconsistency in Platform Analytics scheduled export file types	While scheduling an export for List—Simple visualization, incorrect file types are shown when the property 'com.glide.par.unified_analytics.enabled' is true and 'com.glide.par.v_table_join.enabled' is true or doesn't exist. The supported file types for List—Simple visualization export are PDF, Excel, and Embed List.	Refer to the listed KB article for details.
Analytics Export API PRB2001041	Back button and export functionality don't work with List	List component isn't compatible with Yokohama, so it should be backported. There are also List-related family changes.
Analytics Hub PRB1930155	'Comments created' date inside a widget displays the incorrect date	.
Application Manager PRB1971605	The base version is grayed out in the new Application Manager	The base version is grayed out in the new Application Manager, which stops to install the custom version created on the base version 1.0.0. The new Application Manager allows to install the custom version only on the latest version.
Application Manager PRB1986694 KB2753355	App Manager and My Company Applications incorrectly show available updates after update checker	When publishing a new version of the app, it doesn't appear in 'My Company Application'.	Refer to the listed KB article for details.
Application Manager PRB1992835 KB2804238	A deprecated Classic App Manager URL causes issues on some instances	Users have access to the Classic App Manager by navigating to the URI '/$allappsmgmt.do'. When they switched to new app manager later, and load the 'My company applications' page, that can corrupt the App Manager checksum cache, which causes the manual sync to fail.	Refer to the listed KB article for details.
Asset Management PRB1990771	Asset Analyst UI changes	This is a product update.
Authentication PRB1990715	Token Exchange API for Voice Service Agent	This is a product update.
Change Management PRB1977634	Change ATF test fails/flaps due to ATF steps timeout
CMDB Identification and Reconciliation PRB1957741	On the CMDB Health Dashboard, a parent metric, like completeness, displays a total count of less than one of the sub-metric (required or recommended)	This causes totals to be incorrect.
Content Analytics PRB2002450	JavaScript sandbox replacement for app-content-automation
Content Experiences PRB2002443	JavaScript sandbox replacement for app-content-publishing
Content Governance PRB2002459	JavaScript sandbox replacement for app-content-governance
Contextual Search PRB2010418	An admin user isn't able to add more additional resources to the search context (cxs_res_context_config) table
Customer Operations for Customer Service Management PRB1879390 KB2052806	Viewing history for a csm_consumer can trigger an update to the consumer record	This issue can occur when there's a difference between the address (street, state, zip, country, or city values) on the csm_consumer record and its primary cmn_location. It's caused by the client script 'Set type values' on csm_consumer, which triggers on submit when the form view is 'case'.	Refer to the listed KB article for details.
Data Archiving to Object Store PRB1999471 KB2820738	otherBlobsInChunkExist includes non-storage fields in coordinate comparison, causing premature S3 object deletion	The Archive Destroy job permanently removes archived records after they've aged past a configured retention period. It encounters a software defect that causes it to get stuck in a loop, repeatedly failing on the same set of records.	Refer to the listed KB article for details.
Database Persistence - Data Access PRB1715251	There's a syntax error when adding a dot-walked column with the type 'fx_currency' to a report	The user observes: Syntax - error or Access Rule Violation detected by database ((conne612968) Unknown column 'task1 .amount' in held list').
Database Persistence - Data Access PRB1968109	When glide.db.aggregate. groupby_display_optimize is true, group by reference fields display value is a sys_id
Database Persistence - Data Management PRB1939498 KB2814840	Migration to RaptorDB brings column length changes, which could impede archive table synchronization	The synchronization should run quickly, but this scenario can occur and slow down the process.	Refer to the listed KB article for details.
Database Persistence - Data Scale PRB1970851	Glide enters 'Databaseless' mode after AHA failover is completed	This happens on an instance where the primary server is down.	Open an instance where PG Fastaha is enabled and the primary server is down. Perform AHA failover. Observe that the failover is successful but app nodes aren't coming online. Glide entered 'Databaseless' mode.
Database Persistence PRB1965515	Minimize the impact of connection pool expansion throttling on scheduler jobs
Data Fabric Table Glide Services PRB1990231	Unable to install the Zero Copy Connectors plugin on production due to a conflict in the platform plugin ID	Users have a subscription and are entitled to the Zero Copy Connectors application, but can't install it due to a licensing issue.
Data Fabric Table Glide Services PRB2005124	Plugin ID in LicenseUtil.java should be fixed	LicenseUtil.java checks whether a license for the Zero Copy Connector app exists on an instance using GlideLicenseAPI(). hasLicenseForFeature (APP_ID, ZCC_LICENSE_ID, false, true). Currently, ZCC_LICENSE_ID is set to sn_data_fabric_zcc. This should use the plugin ID instead of the scope name, so sn_data_fabric_zcc needs to be changed to com.sn_data_fabric_zcc.	On a Yokohama instance, try to create connection. Observe that it throws an error.
Declarative Actions PRB1975231	Field decorators for variable fields are cached in a way that causes issues, causing them to not show up sometimes
Discovery PRB1937999	Discovery Schedule should support discovery of IPAM IPs and subnets	Currently Discovery Schedule only supports discovery IP address lists, IP Ranges and subnets configured in associated discovery_range and discovery_range_item_ip tables. This limits the ability to support integration of IPs from other sources such as IPAM.
Discovery PRB1953520 KB2718488	ecc_queue input records are stuck in IP jail, causing status cancellation	IPs are placed in a jail if it is suspected that multiple IPs on a schedule belong to the same device. IPs are supposed to be released from jail when the discovery of the device is complete. When this occurs, the IPs are never released from jail.	Refer to the listed KB article for details.
Discovery PRB1981940	Add global scope support to allow application scope to delete cmdb_file_information records	The cmdb_file_information table is owned by the global scope and requires global-scope handling for controlled bulk deletions.	Create or use an existing application scope. Implement logic in the application scope that requires deletion of records from the cmdb_file_information table using dynamic query conditions (based on runtime parameters). Attempt to perform the deletion directly from the application scope. Observe that the deletion can't be completed due to cross-scope access restrictions and/or the lack of a global-scope utility that supports dynamic, caller-defined queries.
Discovery PRB1996875 KB2801169	Cloud Infra Patterns are launched per datacenter, impacting GCP Discovery Schedules	GCP Discovery schedules take significantly longer to complete because the system launches one pattern execution per datacenter instead of one pattern per service account containing all datacenters. This results in a significant increase in probe executions.	Refer to the listed KB article for details.
Discovery PRB2003122 KB2896586	The Amazon Web Services (AWS) resource inventory box pattern stopped launching after Zurich for a non-US date format	It has a display value that is different from the instance.	Refer to the listed KB article for details.
Discovery Probes and Sensors PRB1960340 KB2611107	Discovery's MID Server script files for Application Dependency Mapping are deleted during clones	The MID Server script file (ecc_agent_script_file) records have attachments. These attachments may go missing after the instance is cloned, breaking Discovery probes. To retain the attachments of base instance code-related records in clones, the table_name in the sys_attachment table must be prefixed with ZZ_YY. Discovery's names don't have that prefix, so the attachments may not be copied depending on the preserver/exclude settings for the clone.	Refer to the listed KB article for details.
Document Intelligence Unified Backend PRB1988914	Add functionality for Input Router to consider an entire document set	This is a product update.
Embedded Help PRB1912461 KB2774125	Default values for properties in an instance must change due to the docs' site URL change	Due to a change in website vendors and their inherent URL structuring, the URLs generated by the in-product help function are broken. Some system properties must be changed.	Refer to the listed KB article for details.
Embedded Help PRB1975389	Add a check in Platform to control the access of foreign language content from Gumbo	Support has been extended to localized content in the Help Center framework. However, content in languages other than English isn't being published to Gumbo. There should be some control in the instance to avoid hitting the Gumbo to access the foreign language content.
Embedded Help PRB2008561	Update the list of allowed domains
Employee Relations Case Management PRB2012556	Users can create more than one active Performance Improvement Plan (PIP) case for the same subject person		Create an employee case for any user with the HR service as PIP. Move the case to the 'Ready' state. 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.
Encryption Support PRB1993181	Flow session cloning (PFSessionClone) needs to load all roles, including snc_internal	GlideSecurityManager. get().getRoles() has side-effect of removing snc_internal role when explicit roles plugin is enabled. Using GlideSecurityManager.get().getRoles(true) will prevent this side-effect.
Event Management PRB1902287	Users can't uninstall sn_em_ai_overview and sn_em_ai_action because they're part of core-ui-components/ 'Core UI Components'/ sn_core_ui_ponents
Event Management PRB1990235	Add an entry point to sanitize an HTML field before insert	This is a product update.
Flow Engine PRB1879472	Sub-flow context can't be opened after an upgrade to Yokohama	There's a deserialize error when viewing a flow context that has a static value assigned to the subflow output. The error stack can be located in the logs.
Flow Engine PRB1944299	'Extension point not found' exception from com.glide.hub.action _step.script plugin	The Send_notifications action executed on a node with new code has no error. The Send_notifications action executed on a node with old code errors out with an exception.	Open a Yokohama or Zurich 2-nodes TD instance. Generate Now Assist Troubleshooting Notifications. Upgrade node 1 to a more recent release. Expected behavior: The Send_notifications action is executed on both node 1 and node 2 without any error. Actual behavior: The Send_notifications action executed on node 1 (the new code) has no error. The Send_notifications action executed on node 2 (the old code) errors out with an exception.
Flows (Family Channel) PRB1990495 KB2780048	The Update record action updates an invalid GlideRecord and triggers business rules when the reference record is deleted while a flow is in progress	When the record that is passed to a subflow is deleted while the flow is in progress, any Update record action in the subflow doesn't error out with 'Glide record is invalid' type errors. Instead, the Update record action in the subflow falls back and iterates on all records in the table, causing substantial data impact.	Refer to the listed KB article for details.
GlideRecord PRB1990713	MID Server core logic for Collector Framework	This is a product update.
GraphQL API PRB1971431	Extremely large tables are queried from GlideRecordSchema, causing out of memory errors	Out of memory exceptions are occurring in the GraphQL framework when GlideRecordSchema is invoked to query tables with extremely high record counts. The cause looks to be from GlideGraphQLQueryHolderImpl .getRowCount returning the rowCount of the GlideRecord when it should be providing the pagination limit size. The value from that method is used to initialize an ArrayList of the provided size in the graphql-java framework ExecutionStrategy.
Horizon Component Library PRB2004216	Update now-avatar properties to receive styling for AI avatars for AI Specialists
HTML Field Type Editor PRB1990982	Predictive text assistance	This is a product update.
HTTP Client PRB1992475	Asynchronous HTTP client upgrade for complete feature change	This is a product update.
Integration Hub PRB1775255	Increase the attachment limit in the SFTP step in Integration Hub	There's a hard-coded value of 100,000 KB for SFTP imports. When the user attempts to import more than the hard-coded limit, it fails. Even though it lets the user to set a limit, if it is higher than 100,000 KB, it isn't allowed.
Integration Hub PRB1989692	Page details are missing from the execution details for Datastream Actions when run via FlowAPI	The 'Page Details' section is missing and the user is unable to view step-by-step execution data (Pagination Setup, Integration Metadata, Step Configuration, API request/response).
Internationalization Features PRB1892286 KB2277705	A non-admin user can't change a dashboard name by specific steps when the system language is set to Japanese	The dashboard name should be updated correctly and reflected in both the primary record and its translated fields, as it is in the Washington DC and Xanadu versions. In the Yokohama version, the update to the dashboard name fails silently when the Japanese language is enabled and the sys_translated record exists.	Refer to the listed KB article for details.
JVM at Scale PRB1967652	Expose microsecond level print of latency to DB into stats/xmlstats for easy bigdata ingestion	This is a product update.
Knowledge Management PRB1976775	A KB's article title is empty under the 'Event property values' column in the 'Article Title' report		Navigate to an instance. Navigate to the User Experience Analytics dashboard. Select Employee Center in the menu. Navigate to Data Foundation > Events section. Open the 'View Knowledge Article' event. Under 'Event properties', navigate to the 'Article Title' and 'Article SysID' widgets. Observe the '...' empty value on the report chart. Export the data to Excel. Observe the exported report contains an empty field under the first column labeled 'Event property values'.
Knowledge Management PRB1981407	After Zurich, the Source task field for knowledge articles isn't assigned	This issue occurs when a user creates a feedback task record for a knowledge article. The feedback task allows knowledge managers to either edit the existing article or create a new one. However, when users attempt to create an article from the feedback task, the new article doesn't reference the originating feedback task as it does in Yokohama. The Source Task field, which should contain this reference, appears in the article form view, but it is grayed out and empty.
Knowledge Management PRB1992747	Now Assist KB creation pop-up is unable to continue with the KB creation when no tasks are available for search	This issue occurs only in Native UI. Using the Create knowledge UI from Native UI during KB creation shows no options available to continue without selecting tasks. This blocks KB creation.
List Administration PRB1894732	When a column is grouped, incorrect sys_ids are returned		Add a record list bundle to the UIB page with groupby Number. Create a declarative action. Preview the page. Select the record. Select the Demo button. Check the console. Observe that the returned sys_ids are improper.
List Administration PRB1991341 KB2803013	Unable to update User field / other fields at Certification Task level	In the Certification Task record, under the inline level, the User field or other editable fields are not editable. When attempting to change the value by double-clicking (or using the keyboard shortcut) on the field, the field goes into a loading state and never allows selection of another.	Refer to the listed KB article for details.
Major Incident Management PRB1960650	There's a slow SQL query on multiple instances when executing the 'ContactManagementEngine .java#L210'	.
MID Server PRB1963685 KB2633251	MID Server File Sync doesn't sync attachments when the sys_attachment table name is prefixed with ZZ_YY to prevent clones deleting the attachments	Attachments of MID Server Synched File records, such as MID Server Script File [ecc_agent_script_file] records, aren't synced to the MID Server when the sys_attachment table name is prefixed with ZZ_YY to prevent clones deleting the attachments. The MID Server platform should be expecting not only the actual table name, but also table name prefixed with ZZ_YY when querying the instance for the attachment.	Refer to the listed KB article for details.
MID Server PRB1966095 KB2659568	Auto MID selection causes excessive metrics to accumulate when connection issues occur	The MID server during auto-mid-selection can experience excessive memory allocation and CPU spikes due to agents connecting with network connectivity issues. As part of auto-mid-selection, agents make REST based calls to the MID server querying for the number of connected agents in order to perform load balancing and honor the maximum number of agents connections allowed. Part of this call will generate a metric instance for tracking which accumulates and can potentially block other thread calls to the MID web server hosting the REST calls. The result is the MID server experiences heavy CPU spikes and excessive memory retention while agents attempt to find the correct MID server to connect with. Coupled with network interruptions, this can result in the MID server running out of memory.	Refer to the listed KB article for details.
MID Server PRB1971767 KB2678137	Agent Client Collector policies aren't synced to agents after MID restart and agent reconnection	The MID server currently has an issue where the policies and their check instances are not associated for tracking if the config_publish ECC queue message has been processed by the MID server before the offline agents reconnect. The information will be present in the policy tracking mechanism, but the flow that processes the message sent to an agent during keep alive processing will not contain the up-to-date policies and their check instances due to the missing associations.	Refer to the listed KB article for details.
MID Server PRB2012665 KB2953193	MID Server upgrades fail because the mid.max.extract .file.size property default is too low	The MID Server property mid.max.extract .file.size property default of 500MB is too low for MID Server upgrades. Upgrades fail at the download/extract stage with the following errors, after the mid-core zip file has been successfully downloaded: '...Exceeds the maximum size (500000000). Unable to download package. The size of uncompressed files is too big or too many entries in the zip file'. This fails before the MID Server is shut down, avoiding outages, but causes repeated failed upgrades. The risks of running with an instance vs MID Server incompatibility, that generally has been known to break some features and cause data loss.	Refer to the listed KB article for details.
Mobile Experience for Field Service Management (Glide Family) PRB1978601	Offline Mode fails with an error due to FSM Data Item query fetching assets even if the Location field is empty		Configure the Field Service Mobile plugin. Have many records in wm_task where the 'Assigned to user' has their Location field empty, which will be returned by the 'DI_my_inventory_remove_parts' data item. Log in to Agent Mobile with FSM active. Enable Offline mode. Attempt to download cache. Notice that an error occurs.
Mobile Platform PRB2006346	The offline V2 sync API returns cumulative table snapshots in each execution result	This issue has been present since Yokohama or earlier and affects the Mobile Agent during offline-to-online sync. When a user performs multiple record edits in offline mode and then reconnects, the sync API processes each queued action and generates a RefreshedDocument per execution result. Due to the cumulative snapshot behavior, each execution result incorrectly includes the data of all previously synced records — not just the one modified in that action. This causes the response payload to grow at an N² rate: N offline edits produce N² total records across all execution results. The impact is significant. Users who perform hundreds of edits in offline mode can experience sync payloads exceeding 100MB+, leading to severe performance degradation, potential sync failures, and a poor user experience at the moment of reconnection.	Provision an instance with Field Service Management with demo data installed. Log in to the agent app with a field agent user. Navigate to the offline settings. Download the offline cache. Go offline. Open a work order task form screen and make an edit. Open a separate work order task and make an edit. Navigate to settings. 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.
Natural Language Query (Family Release) PRB1992648 KB2794876	com.snc.nlq.MetadataGenerator modifies GlideElement objects in TableDescriptorCache	This issue is observed in production.	Refer to the listed KB article for details.
Normalization Data Services PRB1931854 KB2664685	A core company isn't handled correctly for a domain-separated environment	In a domain-separated environments with Normalization Data services turned on, creating models or configuration items may fail or use an incorrect manufacturing (publisher) company. When selecting a manufacturer from a specific domain, the system incorrectly replaces it with a company record from the user's domain.	Refer to the listed KB article for details.
Now Assist Panel PRB1985088	Link in NASS doesn't open in a new tab		Navigate to full page. Search 'What is spam?'. Select the source link for the KB. Expected behavior: The KB link opens in a new tab. Actual behavior: The KB link doesn't open at all.
On-Call Scheduling PRB1981295	OnCallRotation. getPrimaryUser returns null when there's no members on a roster and only coverage present	.	Log in to instance. Make sure an assignment group has the roaster with no members and has only the Primary Coverage. Execute the script. Expected behavior: The user id is returned. Actual behavior: The result is null.
OneExtend PRB1981515	Add enhanced debug logging around async client call	Debug logs need to be added when each streamed chunk is received, when the final callback is received (VAStreamConsumer), before enqueueing in the hybrid queue, and after enqueueing in the hybrid queue and the callback is complete.
OneExtend PRB1997461	NowLLMStreamReader.parse() splits on 'data:' inside JSON content, corrupting LLM stream output	NowLLMStreamReader.parse() uses data.split'data:' to split SSE (Server-Sent Events) frames. This splits on all occurrences of the literal 'data:' — including those embedded inside JSON string values in LLM output. When the LLM produces certain content, the split fragments the JSON into unparseable pieces. Jackson silently fails (catches Exception in extractResults), content is lost, and the feature invocation stays stuck in 'processing' with null output.
OneExtend PRB1998199	NAVA gets stuck intermittently		Ensure Response Streaming is enabled. Open NAVA. Enter 'Creator.' Notice that NAVA will get stuck intermittently.
Performance Analytics PRB1948103	After upgrading from Xanadu to Yokohama, the line chart appearance changes for the 'current year vs. last year' comparison report using data sets	This fix requires manual change of a report property. Set 'Allow data label overlap' to true from sys_report record. After having patch for this feature, the customer needs to update this property for each required report manually. This can be done from platform list by adding column 'Allow data label overlap'.	Open any Xanadu instance. Create a Line report on sys_trend with the following: Trend on: Created Condition (Dataset A – Current Year) Add a second dataset with the same configuration but for last year. Condition (Dataset B – Last Year) Save the report. Observe the chart appearance in Xanadu. Upgrade the instance to Yokohama. Open the same report. Observe the chart appearance again.
Platform Analytics Component API PRB1984668	There's a column filter issue in the Data Visualization library page, in which the 'Does Not contain' isn't working on the 'Owner' column	Applying the 'Does not contain' filter on the 'Owned by' column does not filter the records properly.	Navigate to the Data visualization library page. Apply the 'Does not contain' filter on the 'Owned by' column. Expected behavior: All the records should be filtered by the applied filter value. Actual behavior: It is not filtered properly when there are multiple values.
Platform Analytics Dashboard API PRB1843154	Not able to save a dashboard after editing
Platform Analytics Dashboard API PRB1846141	Certain missing properties for a visualization can cause side effects	Some impacts include drilldowns in a dashboard not working, and the Save button being enabled without any edit.
Platform Analytics Dashboard API PRB1969752	Cached toolbox items cause options to remain in a previously selected language	While working on the Platform Analytics Dashboard and Data Visualization modules, users have observed that the language in the right-hand menu changes automatically without any user action. The issue occurs intermittently. Sometimes, it appears only in a single browser session, while at other times it affects multiple browsers simultaneously. No consistent trigger has been identified so far. Users report that refreshing the page or clearing the cache may temporarily resolve the issue, but it can recur unpredictably.
Platform Analytics Dashboard API PRB1983811	Unexpected response format for users without assigned roles	The API should return properly formatted JSON with error code, but instead it returns an invalid format.	Run GetMobileVisualizationIT #verifyAsNoRoleUser. Expected behavior: The API returns properly formatted JSON with error code. Actual behavior: The API returns an invalid format.
Platform Analytics Dashboard API PRB2000628	DashboardWidgetDao.update() does not persist widget_props field	When updating a dashboard widget record via DashboardWidgetDao.update(), the widget_props field is not written to the par_dashboard_widget table. The create() method correctly calls gr.setValue(WIDGET_PROPS, widget.getWidgetProps()), but the update() method is missing this call. This causes widget-level property overrides (for example, followFilters, header color, header title) to be saved on initial widget creation but silently lost on any subsequent dashboard save.
Platform Analytics Migration API PRB2003609	After migration, widget settings are lost	All of the widget related properties (follow filter, header settings, etc) should be migrated to the widget_props of par_dashboard_widget table.	Create a CoreUI dashboard. Add a report widget that's used in other dashboards and set the header alignment to be center. Migrate the dashboard. Verify that the par_visualization record is created for the sys_report. Open the dashboard in PAE. Expected behavior: The header alignment settings are migrated. Actual behavior: The header displays with the default settings.
Platform Runtime PRB2014434	The Granular Delegation API isn't returning any results
Playbooks (Family Channel) PRB1984620	Loading of sys_pd_process_definition_ c5142dddff37e21082a8ffffffffff2d taking two hours	This issue occurs when upgrading Yokohama and Zurich.
Playbooks (Family Channel) PRB1991515	After a restart, the stage is still displayed even when the conditions aren't met	Remnants from the previous playbook lanes remain visible, although they appear in a 'PENDING' state. However, those lanes are reached, they update correctly.	Navigate to Configurable Workspace > Lists > AGR Connections > All. Create a record with subcategory 'WebEDI'. Proceed with Playbook. When asked for 'Fast Follower', select True. Proceed with Playbook until 'Packaging Data Exchange' (lane 7) is reached. Perform the complete Playbook restart. Proceed with Playbook to 'Validate Connection Type' (activity 1.4) activity. Change the subcategory to 'ClassicEDI'. When asked for Fast Follower, select False and proceed. Proceed with Playbook until 'Testing' (lane 9) is reached. Perform the complete Playbook restart again. Proceed with Playbook to 'Validate Connection Type' activity. Change the subcategory to 'WebEDI'. When asked for Fast Follower, select False and proceed. Observe that the 'Fast follower' phase is initiated even if the fast follower is not marked, which isn't correct. Also, observe that even if the subcategory is 'WebEDI' now, the phases / activities for 'ClassicEDI' are still initiated, which isn't correct. Expected behavior: If fast follower is set to true, then WebEDI and ClassicEDI don't appear on the playbook. If fast follower is set to false, then WebEDI or ClassicEDI appear on the playbook. Actual behavior: If fast follower is set to true, WebEDI and ClassicEDI appear on the playbook. If fast follower is set to false, then fast follower, WebEDI, and ClassicEDI appear on the playbook.
Playbooks (Family Channel) PRB1992485	Remove sys_ids from Playbook-fetched telemetry metrics	Process definition information isn't included as part of data that is captured.
Playbooks (Family Channel) PRB1999600	Playbook with questionnaire activity is not activated by default in Yokohama instances	When apps are uploaded from Nexus or from a zip file, and they contain playbooks with questionnaire activities, those playbooks are missing trigger data. When a playbook has a questionnaire activity, a line gets added to the XML that causes everything after that line to fail to load.
Schedule Optimization (Glide Family Channel) PRB1988877	Fix default skill level logic for Schedule Optimization 1.1	An additional validation is needed to ensure that the default skill level is returned correctly for skills.
Search Administration PRB1968215	Elements must only use supported ARIA attributes for the Search catalog field for Service Portal (SP)		Ensure there are multiple requests raised in SP. Launch a base instance. Navigate to Request on the header menu and activate it. Notice that multiple records are displayed. Choose any request. Open the Request summary. Navigate to Search catalog. Notice that menu items are displayed. Run 'Evinced Flow Analyzer' and AXE devtool. Observe the issue. Expected behavior: The ARIA attribute is not allowed as seen in 'aria-selected='false'' and an element's role supports its ARIA attributes. Actual behavior: Elements only use supported ARIA attributes, as seen in the element location: sn-search-combobox, sn-search-combobox-desktop,. -search-popover__item--is-active.
Service Catalog Builder PRB1982503	In Yokohama, the Continue button doesn't work and the Preview button is missing in the canvas view of t2c- 6.0.7		Open a Yokohama instance with NAFC 27.6.1. Enable the T2C skill. Navigate to Catalog Builder. Select Build with Now Assist. Enter any utterance. Observe that the Preview button is missing and the Continue button doesn't work.
Session Management PRB1962465	Performance degradation with build 1108
Software Asset Data Import PRB1975879	PPN lookup shouldn't be performed for empty PPNs	If there is a corrupted PPN record where the part number value is empty, it will be returned and set in the PPN reference value.
Software Asset Data Import PRB1993170	Pass the auto-incremenator using the unique ID of the table rather than using the contract number	There should be automation for both scenarios: 1. Uploading the docs, finishing the flow, and importing another doc to the same contract. 2. The asset covered from both docs should be part of the same contract.
Software Asset Data Import PRB2001908	Rename the Ignore button to 'Exclude'	This is a product update.
Software Asset Data Import PRB2001911	Update text on the content opt-in page	Software entitlement details are extracted from software contracts using AI. Opt-in to the content service to help improve the prediction for product, publisher, and edition values.
Software Asset Management PRB1936688	Add-On Publisher values are not populated properly	When Content is shipping Add-On Publisher life cycles, it gets defaulted to Publisher value instead of Add-On Publisher.
Software Asset Management PRB1975292	Contract start date parsing doesn't work as expected	The issue occurs when alphanumeric input is given. The Organization ID field should be set to read-only after successful validation.	Open the docusign integration profile form. Set Integration type = Org Level. Set the contract renewal date to any number. Observe that the contract renewal date is checking for integers, but the functionality is different in classic and workspace. In classic, any alphanumeric value is reduced to just integers. In workspace, the value stays as-is and a warning message appears. Open the DocuSign integration profile form again. Set Integration type = Org Level. Populate the fields. Save. Validate. Change the org ID. Save again. Observe that the Organization ID field isn't set to read-only after successful connection validation.
Software Asset Management PRB2001199	Trigger flow and alternate flows	This is a product update.
Time Card Management PRB1988000 KB2760220	Managers are no longer able to add time cards for their direct reports and delegates in the Time Sheet Portal		Refer to the listed KB article for details.
UI Field Administration PRB1962435	There's an incident form load issue due to the missing macro sn_wwna_nacm_component, which causes a performance issue	When loading an incident form, the local host log is consistently flooded with the 'DBMacro WARNING * WARNING * Macro missing: sn_wwna_nacm_component', and this is causing the form to load slowly.	Open any incident record in the instance. Navigate to System Diagnostics > Session Debug > Enable All. Observe the error in the localhost log. The warning is called many times: 'DBMacro WARNING * WARNING * Macro missing: sn_wwna_nacm_component'.
UI Field Administration PRB1990980	Add backend support for the multi-section configuration and fetch configurations in the UI to create the BE layer	This is a product update.
UI Field Administration PRB1990983	Support core IT use cases	This is a product update.
UI Form Administration PRB1920717	Hyperlinks in annotations on forms are not opening in a new window	Hyperlinks open in the same tab instead of opening in a new tab.	Open a base instance. Open an incident table. Open any incident. Select and hold (or right-click) and navigate to Configure > Form builder. Create an annotation with the code. Open the incident in SOW. Under 'Details,' select the link. Expected behavior: It opens the link in a new tab. Actual behavior: It opens the link in the same tab.
UI Form Administration PRB1959255	g_form.clearValue shouldn't clear a choice field if the field is empty or has the default value	In Service Operations Workspace, the onChange client script only runs in the workspace if a reference isn't on the layout.
UI Form Administration PRB1971158	A blank screen displays when a user tries to access the ast_contract record present in another domain	When the user doesn't have access to the record, globals are coming as undefined, and aren't adding the default value of {}. This causes a JS error later in the processing, which results in the blank screen.
UI Form Administration PRB1972511	Invalid payload event for CTRL_RECORD# MACROPONENT_PROPERTY_CHANGED	There are properties that are defined in the CTRL_RECORD# MACROPONENT_PROPERTY_CHANGED event definition. The properties are not stored in the cached payload due to a parsing error. getProperties returns no properties due to a parsing error.
UI Form Administration PRB1993929 KB2908761	RCAs are created with the source as 'sys_ws_operation.do? sys_id=6986d96dff7322102 ae8ffffffffff48' for HR apps	An RCA is generated: 'RCA triggered - 'Read operation on table 'sn_doc_html_template' from scope 'Global' was denied. The application 'Document Templates' must declare a Restricted Caller Access privilege. Please contact the application admin to update their access requests'. None should be generated.	Refer to the listed KB article for details.
UI Form Administration PRB2004264	Updates to support AI Specialist and draft note	This is a product update.
UI Form Administration PRB2011302	An API /api/now/ai_agent/modified_fields call on a form load increases the browser time	An API call has been added for every form operation as part of the platform engine feature '/api/now/ai_agent/modified_fields'.
Usage Analytics PRB2002024	There's an inaccuracy in IH transactions DEFN ID	The 'IH Usage' table has an accrual month logic, where the same record is updated every single day for the rest of the month. The IH transactions fabric credits DEFN ID only fetches the first day reporting of usage on a record, and not the rest of the 29 days.
UXF Macroponent PRB1899084	The technical dashboard override isn't applied when switching domains in UI Builder (UIB)		Open an instance with the 'Domain separation' plugin installed. Create a technical dashboard within the global domain. Add a data visualization to it. Switch to another domain. Create a page override. Add an additional data visualization to the overridden page. Preview the created dashboard pages from UIB by switching domains. The page content changes as expected. Open the technical dashboard via the Platform Analytics experience. 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.
UXF Macroponent PRB2009619	Multiple stuck threads cause semaphore exhaustion, requiring a node restart	There's a stuck semaphore transaction related to the function createCacheablePageFragmentFromShell in CacheablePageFragmentFactory.java. This can lead to multiple threads being stuck with no option but to restart the node to provide relief.
UX Framework PRB1947452 KB2601019	UX screen record allows the Parent Macroponent field to be set to the screen within the same screen collection as the current screen record	This results in an invalid screen configuration.	Refer to the listed KB article for details.
UX Framework PRB1966630	Workspace tabs aren't rendering screens	Loading a record page via select through (hydrate) is missing the audience applicability check (present in direct load), leading to incorrect screen variant resolution.
UX Framework PRB1990140	Framework support for the AI screen reader Store app	This is a product update.
Versatile Node and Cluster Configuration PRB1819117	Upper case datacenter name results in no triggers being executed	DNS PTR (reverse-DNS) records return hostnames with uppercase 'ABC123' instead of the expected lowercase 'abc123'. Comparison checks in HostnameProximityEstimator should be case insensitive.
Versatile Node and Cluster Configuration PRB2002618	There's an AHA check failure due to a node_id mismatch from capitalizing the host name	Domain Name System (DNS) PTR (reverse-DNS) records can return host names with uppercase instead of the expected lowercase. The HostUtil.getHostname() should return lowercase.	Start a node with a set system_id. Note the node_id. Modify the system ID to have 'AMS' capitalized. Restart the instance. Observe that the node ID is now different. Modify the system ID back to the original value. Restart the instance. Observe that the node ID is now back to the original.
Virtual Agent PRB1916016	The End Conversation action does not end live agent interaction when the Post Chat survey is enabled	The interaction and conversation should be closed once the user responds to the survey.	Start a conversation with a Live Agent using VA API. Make sure the Post Chat survey is enabled with the 'Requester Ended' flag. Exchange messages between the agent and user. Initiate the END_CONVERSATION action from the user. Expected behavior: The interaction closes when the user ends the conversation, then the survey is presented to the user. This behavior can be verified from webclient as well. Actual behavior: The interaction is kept open and the survey is presented to the user.
Virtual Agent PRB1919055	Add a new property to turn off the 'shorten responses' feature by default	After upgrading Xanadu, shorten responses were added and set as 'on' by default. There is no method to set up shorten responses to 'false' as default. The user cannot see the full test display as default as before.
Virtual Agent PRB1960339	The Use an AI agent action moves to a 'Waiting' state when called inside a flow-logic component	In a use case where calling the agent multiple times in a single subflow/flow using the flow-logic, the agent is initially called, but the second call goes into 'Waiting' state and no other calls go through. In the second call to the same agent (with a different objective), the execution record's state is 'Ready' and the message 'Conversation data is not populated' occurs for the second record. This leads the action to be in 'Waiting' state. The sysevent record is also not created.
Virtual Agent PRB1973526	The business rule 'VA Phone Interaction - On Close' is not firing in the 'wrap_up' state	The outcome of this issue end to end is this is causing an interaction that is intended to be closed by the agent (so has entered wrap up) resurfacing when the same caller makes consecutive calls before wrap up is finished.
Virtual Agent PRB1978979	The session should not get timed out during the execution of topic as a tool when run as an AI User	The user session is getting timed out after uploading the attachment.
Virtual Agent PRB1982735	Handle an unknown document_id field in the sys_cs_provider_application		Set up Glide NLU or Now Assist. Set up the appropriate Teams version. Start a conversation from a bot. Logs have an unknown field document_id in sys_cs_provider_application exception for every Teams request.
Virtual Agent PRB1984195	Interactions are reused every time the startVoiceInteraction Voice API is invoked	The API should create an interaction, but instead it reuses the interaction if it's active. This used to work correctly in Yokohama, but is now failing.
Virtual Agent PRB1985280	Duplicate attachment is displayed in Virtual Agent after portal refresh		Open an instance in two browsers. In browser one, impersonate the agent 'Fred Luddy'. Set the agent as available in Service Operations Workspace. Switch to browser two. Impersonate as admin. Initiate a chat from the portal or web-client. Connect to a live agent. In browser one, accept the chat as the agent. Send any image as end-user to agent. Refresh the portal at end-user. Expected behavior: Nothing happens. Actual behavior: The duplicate attachment/image is displayed only in Virtual Agent.
Virtual Agent PRB1989732	The assistant_personalization system property should handle default case properly for branding and personalization	When updating the sn_nowassist_va. assistant_personalization system property to a random value, it doesn't behave the same as the default value.	Open the sn_nowassist_va. assistant_personalization system property. Update the value to a random value, different from the given three choices. Run a query in the Virtual Agent portal connected to the assistant that has personalization. Navigate to the sys_generative_ai_log table. Locate the record for either AIA Unified Planner, AIA Unified Planner Quick, or AIA Planner. Review the record and verify that the configured persona is appearing in the prompt, in between <persona> and </persona>. Expected behavior: It behaves the same as the default value, and should return only agent_persona from the configuration attribute table. Actual behavior: It behaves as TONE_RESPONSE_LEN_PERSONA and returns all the details.
Virtual Agent PRB1990179	Language Detection from French to English doesn't show the topic name in English
Virtual Agent PRB1992132	The Virtual Agent (VA) link menu items are not displayed correctly	In Enhanced chat, VA link menu doesn't show the labels the way it shows on the standard chat. Instead it shows the label as plain text and shows the link below. If the link is longer, it increases the width of the 'Contact support window'. It should only show the label and shouldn't show the link.	Open the Menu items from the branding. Add a long link. Notice that the width of the window increases beyond the chat window, and the label is shown in plain text and the link shows.
Virtual Agent third-party integrations PRB1980510	Allow linking for AI agent user type from AI agents app	When the user triggers auto-linking from AI agent with an AI agent type, the linking fails.
Virtual Agent Web Client PRB1981024	Markdown link [label](url) renders an incorrect URL when the URL contains a dollar sign ($)
Virtual Agent Web Client PRB1981038	There's an extra empty message added after the HTML output when messages arrive too quickly	This was added for scrolling. When the messages come too fast, this 'hidden' message stays in the chat.
Virtual Agent Web Client PRB1983355	Markdown to HTML parsing issues when there are special characters in the payload		Create a topic in the designer with a text node with an response message containing a ` (backtick) character. Test the topic. Notice the response shows unintended HTML. Expected behavior: Parsing is broken and links are highlighted. Actual behavior: The HTML output is broken.
Visual Task Boards PRB1994303	In a Visual Task Board (VTB), the 'Select lane' menu displays lane values from all the boards available irrespective of the board selected to move
Word Document APIs PRB1967960	Custom fonts in Word documents throw an error		Identify contract documents with custom fonts embedded into XML. Upload the contract into the contract request. Note that the upload fails and the file isn't visible in the contract request.
Word Document APIs PRB1987850	Issue with document sync when paragraph properties do not exist	An error occurs when attempting to sync the document. This issue occurs when the document was generated using ChatGPT or other AI tools, or when content controls are positioned on blank text.	Make changes to the content control fields with 'Track Changes' on. Observe that the selected paragraph should not have any styling. Attempt to replace the value. Notice that it will throw an error.
Work Order Management PRB1917452	Travel time and work time validation doesn't prevent the work order tasks from being updated, and the Close incomplete button has unexpected results	Modifying the Actual travel duration to be larger than the time between Actual travel start and Actual work start triggers a business rule to display 'Total travel time must be before actual work start' and abort the change. However the change is not aborted. Moreover, if the change is saved with the button Close incomplete, the task status will be updated to 'Closed complete'.
Zero Copy Connectors (Glide) PRB1984931	The WDF table throws exception when referencing glide tables with long logical table names that are greater than 30 characters	An error occurs when opening the WDF table and including the reference column in the list view.	Create a glide table with more than 30 characters. Create a WDF table with a column that references a column in the glide table. Open the WDF table and include the reference column in the list view. Notice than an error shows on the page.
Zero Trust Access PRB1976242	There's a security constraints access prevention issue after enabling zero trust access when opening a VTB dashboard		Provision an instance with the 'Zero trust access' plugin installed. Enable the ZTA property (glide.authenticate. session_access.enabled). Navigate to vtb_board.list. Open any dashboard having a condition with a date-time filter. Select Show board under related links. Expected behavior: The VTB board should load and the session should remain with the roles the user is entitled for that session. Actual behavior: Users observed a security constraints access 403 error for any subsequent action.

Fixes included

Unless any exceptions are noted, you can safely upgrade to this release version from any of the versions listed below. These prior versions contain PRB fixes that are also included with this release. Be sure to upgrade to the latest listed patch that includes all of the PRB fixes you are interested in.