Turbine On-Prem 26.2.0 Release

turbine 26 2 delivers significant enhancements in automation development and platform management the release improves solution deployment through enhanced import/export capabilities for application records developer productivity gains include a redesigned transform data native action, configurable payload handling for webhooks and http actions, canvas annotations, and numerous canvas usability improvements overall, turbine 26 2 0 enhancements make turbine more accessible, more intelligent, and easier to use day to day, helping teams deliver automation faster, onboard new users more quickly, and spend less time navigating the platform and more time driving operational outcomes what's new in this release? hero ai credits and usage dashboards turbine 26 2 0 adds hero ai credits , hero ai prompts , and hero ai tokens under admin panel > usage , replacing the legacy hero ai usage page all accounts can open these dashboards to review hero ai consumption by playbook, component, user, and token category the hero ai credits dashboard includes a credits summary with contract month totals when the hero ai credits feature is enabled for your account monthly credit allocation, remaining credits, and usage progress apply only to accounts on the hero ai credits pricing model hero ai prompts and hero ai tokens remain available for usage visibility when no credit contract applies administrators can filter usage by tenant, date range, and automation type accounts on credits based pricing can configure usage caps and threshold notifications to manage consumption across hero ai features the legacy hero ai usage page is fully deprecated in 26 2 0 when you use the new dashboards to review data from before the release, keep the following migration limits in mind hero ai prompts and hero ai credits — filters such as playbook , component , and model type rely on fields that were not captured in older usage logs those filters do not apply to historical data from before 26 2 0 ; the tenant filter continues to work across all dates for older date ranges, the detailed usage table at the bottom of the dashboard may appear empty even though summary bar charts (total prompts or credits, average per day) still display correctly hero ai tokens — token counting is not retroactive the hero ai tokens dashboard shows usage accumulated only after the 26 2 0 release playbook generator enhancements hero ai now provides a more guided playbook building experience when additional information is needed to generate a playbook, hero ai asks targeted follow up questions with selectable options, allowing you to provide context directly in the chat before generation continues while playbooks and components are being created, hero ai displays progress updates that show what the playbook generator agent is currently working on this provides greater visibility into the generation process, especially for complex, multi flow playbooks build widgets with hero ai hero ai now supports widget building mode within application making it faster and easier to create and customize widgets describe layout changes, code updates, or desired widget behavior in plain language using the build with hero ai option hero ai automatically updates the widget code, allowing user to preview the results before applying and saving your changes the widget editor now opens alongside the hero ai companion in a single workspace, enabling user to edit code, preview outputs, and refine prompts without switching between screens these enhancements help orchestrators to create and iterate on custom widgets more efficiently hero ai native action model selection the hero ai native action now supports model selection on the prompt tab when configuring a hero ai native action in a playbook or component, you can choose from available aws bedrock models instead of relying on the default platform routing alone model selection applies to that action configuration and helps teams align automation with their approved model tiers import/export application records in ai soc the ai soc solution now supports exporting and importing application records for supported ai soc applications (knowledge base articles and routing rules) use this capability to package records with application configuration when moving ai soc solution content between development, test, and production environments during export, optionally include records and filter which records are packaged with the condition builder during import, configure how records are created and matched in the destination tenant this enhancement reduces manual effort when deploying ai soc knowledge base and routing rule content across tenants it is intended for ai soc solution migration workflows install connectors, components, and ai agents directly from canvas canvas now allows users to discover and install connectors, components, and ai agents directly from the add panel while building playbooks users can browse available content, view details, and install or connect content without leaving the canvas installed connectors immediately make their associated actions available in the actions panel, while installed components and ai agents become available in their respective tabs for drag and drop use in playbooks search, filtering, and sorting options help users quickly find the content they need this enhancement streamlines the playbook building experience by keeping content discovery, installation, and workflow design in a single location, helping users build automations more efficiently improve connector asset configuration and validation experience connector asset management has been enhanced with built in testing and validation capabilities users can now test connector assets directly during configuration from the asset details page, allowing connection issues to be identified and resolved before assets are used in automation workflows assets can be tested individually from both the tenant asset page and config manager config manager also supports scheduled bulk validation, enabling organizations to automatically test connector assets on a recurring basis, such as through nightly validation jobs in addition, users can view test results, execution history, audit logs, related playbooks and components, and associated playbook runs directly within the platform tenant deletion and status visibility tenant management has been enhanced to provide greater visibility and control throughout the tenant lifecycle account administrators can now delete tenants they have created directly from the platform tenant deletion is permanent and cannot be undone the tenant management interface now displays clear status indicators, showing whether a tenant is being created, is active, is being deleted, or has encountered a failure these updates improve transparency during tenant operations and help administrators track progress more effectively canvas annotations canvas now supports annotations, allowing users to add documentation and context directly within solution and component editors users can create, edit, move, show or hide, and delete annotations on the canvas annotations support standard editing actions, including cut, copy, paste, and multi select operations, and remain available across editing sessions to improve collaboration and maintainability annotations bring documentation and context directly onto the canvas in solution and component editors enhanced run tab experience the canvas run tab has been redesigned to provide more space and visibility when testing and troubleshooting playbook the updated experience includes an expanded run view, improved test console accessibility, and a streamlined layout that makes it easier to navigate flows, review execution details, and debug playbook runs cancel active playbook runs you can now cancel active or queued playbook runs directly from the platform instead of waiting for a run to finish on its own when you expand run details for an active or queued run, a cancel run button appears in the run header use it from the canvas test console while testing a flow, or from operational health > runs when monitoring production executions after you request cancellation, the run status updates when the engine stops the run this helps you recover quickly from long running tests, stuck runs, or playbooks you started by mistake automatic focus for new properties when creating a new property within an action, the name field is now automatically focused this removes an extra click during configuration and helps users build playbooks more efficiently when working with multiple properties undo and redo support in configuration dialogs undo and redo capabilities have been extended to configuration dialogs throughout canvas users can now undo and redo changes within action, trigger, component, and http configuration dialogs using both on screen controls and keyboard shortcuts this enhancement provides a more consistent editing experience and makes it easier to safely modify complex configurations canvas webhooks and http actions support configurable payload processing webhooks and http actions now support configurable payload processing options users can explicitly define how request and response payloads are processed instead of relying on automatic detection this provides greater control over data ingestion and improves support for payload formats that may not be handled correctly through automatic detection alone redesigned transform data native action the transform data native action has been redesigned to match the modern canvas action experience and make data transformation easier to configure, test, and maintain each transformation block now uses a structured layout with dedicated inputs, transformation, and test areas you declare block inputs explicitly, run tests to inspect results, and discover output types so downstream actions are not limited to generic dynamic labels at design time an outputs tab lets you review discovered outputs and mark sensitive values the redesign also adds a visual transformation builder, optional advanced jsonata editing, block reordering, and improved guidance when selecting transformation types these updates help builders work faster with one of the most commonly used actions in canvas playbooks reference jsonata errors as playbook properties transform data and other jsonata based actions now expose block level error output as mappable playbook properties, similar to script action error outputs when a transformation block fails — including when an advanced jsonata expression uses $error or $assert to return a custom message — the error is stored on the playbook run and appears in the action error output for that block you can reference the message (and related error fields) from downstream actions when mapping inputs, without copying details manually from the run log re designed charts for dashboards and reports turbine now integrates an updated charting engine for dashboards, reports, and related visualizations the new rendering engine improves chart performance and visual consistency across chart types, including bar, line, pie, donut, and grid layouts tooltips, legends, sorting, and color schemes have been refined to provide clearer data presentation and a more consistent experience when building and viewing charts this update is delivered through the enhancedcharts feature flag when enabled, existing dashboards and reports use the new chart components while preserving current chart configuration options record event test data shows recent records first when you select test data for a record event trigger in the playbook test console, the record list now shows the most recently created records first this makes it easier to test playbooks against current records without scrolling through older data git improvements – ui/api timeout errors during git publish git integration has been enhanced to improve reliability during content library publishing operations error handling has been improved to provide clearer messages when publish operations encounter lock timeouts, making issues easier to identify and resolve additionally, git api operation status is now tracked to improve troubleshooting and support more resilient publishing workflows classic playbooks self service conversion tool for classic playbooks to canvas style turbine now includes an in product conversion tool that helps customers migrate classic playbooks to canvas playbooks and components the feature is behind the feature flag if you are interested in converting please reach out to support team users can initiate conversions directly from the ui and track progress while the conversion runs in the background the conversion process handles common playbook structures and dependencies, helping accelerate migration to canvas while reducing manual effort known limitations disabled playbooks will not be converted playbooks with recursion are not supported and will not be converted playbooks with a webhook trigger but no webhook url will not be converted notes all playbooks will be disabled after conversion enable the required flows to begin testing if a classic playbook includes a playbook button trigger, it will be removed from the classic playbook and added to the canvas playbook after conversion addressed issues records record saves succeeded when workflow required fields were empty fixed an issue where a field marked required by a workflow condition still saved successfully when left empty if you had not clicked into or focused that field first, even though the record ui showed the field as required record save now validates workflow required fields on submit without requiring field interaction first record view failed when a large text field file was missing fixed an issue where opening a record failed when a large text field referenced a file that no longer existed or was not found the record view now opens and handles missing large text content without blocking access to the rest of the record calculated fields stripped quotes from text values fixed an issue where calculated fields removed quotation marks from text field values in the calculated result, especially when the quoted value appeared at the end of the formula output calculated results now retain the quotes present in the source text values bulk edit select all failed with grouped text filters fixed an issue where select all in bulk edit did not update any records when your filter used grouped conditions that included a text field equals comparison in a nested group bulk edits now apply to all selected records for those filter combinations csv import failed when similar field names were present fixed an issue where re importing previously exported csv data failed when the file contained fields with similar names (for example, header form and headerform ) csv import now resolves those columns correctly and completes successfully deleting playbook buttons in a copied application affected the original application fixed an issue where removing a playbook button from a copied application also removed the button from the source application and removed the button trigger from the associated playbook deleting buttons in a copy now changes only the copied application disabled playbook button stayed clickable for non administrators on records fixed an issue where the playbook button on a record stayed enabled for non administrator users when the associated playbook was disabled, and clicking it returned a system error the button now appears disabled for all users when the playbook is disabled playbook record actions failed when underlying record data was invalid fixed an issue where playbook record action updates failed or left records in a bad state when underlying record data was invalid playbook updates to affected records now complete successfully record comments were lost during concurrent updates fixed an issue where comments disappeared when a user added or edited comments on a record at the same time a playbook or another user updated the same record comment updates now merge safely without overwriting other changes calculated field formula validation returned inconsistent results fixed an issue where validating the same calculated field formula returned valid in one check and invalid in another for an unchanged formula in complex applications formula validation now returns consistent results for the same formula and field references reports and dashboards line and area charts showed no data with default group by on large data sets fixed an issue where line and area charts in reports displayed no data when a timestamp field used the default group by setting and you had not selected hour, day, week, month, or year, especially with large numbers of records charts now render data with the default group by selection report and dashboard actions in audit logs showed the wrong category fixed an issue where create, read, update, and delete actions on a report appeared in audit logs under record instead of report, and the same actions on a dashboard appeared under record instead of dashboard audit log entries now show the correct category for report and dashboard activity dashboard report cards ignored time filters on first load fixed an issue where report cards on a dashboard showed all records on first load even when the underlying report used a saved time filter, requiring you to set the filter manually after each page refresh dashboard cards now apply saved time filters on initial load report column order changed when toggling column filters fixed an issue where selecting or clearing columns under column filters in a report reordered columns in the data table away from your preferred layout column order now stays stable when you change visible columns dashboard report cards lost pagination when resized tall fixed an issue where report cards on a dashboard taller than four layout units no longer showed pagination controls, making it impossible to move through all results in the card pagination controls now appear at the expected card sizes dashboard date range filter was off screen for read only users fixed an issue where the date range filter on a dashboard was drawn outside the visible area when you had read only dashboard permission and did not have add a card the apply control was outside the visible area, making it impossible to run the filter the date range filter and apply now stay visible and usable browser console showed repeated errors related to in app notifications fixed an issue where the browser console logged repeated errors related to in app notifications while you worked in playbooks, dashboards, components, or application records, or when you switched tenants in app notifications continue to work as expected, and those console errors no longer appear report date only filters returned records off by one day fixed an issue where report filters on date only fields using absolute comparison operators (equals, greater than, less than, and similar) shifted the selected date back one day and returned incorrect records when you reopened the condition builder date only filters now match the date you selected and return the expected records dashboard date picker did not highlight the selected range after reload fixed an issue where the date picker on a dashboard did not visually highlight the active date range after you reloaded the page, even when the filter was still active the selected range now appears highlighted when you reopen the date picker dashboard date time filters behaved differently for relative and absolute ranges fixed an issue where relative ranges (for example, last week) used full day boundaries but manually selected absolute dates used the current time, producing different results for equivalent periods relative and absolute selections now use consistent time boundaries report save failed when grouping by reference fields fixed an issue where creating or updating a report that grouped by reference fields failed to save and returned an error reports with reference field grouping now save successfully scheduled dashboard reports timed out or failed to generate fixed an issue where scheduled dashboard reports timed out, failed to generate card content, or returned system errors when you tried to download a preview scheduled dashboard reports now complete and deliver the expected card output reference field values were missing in paginated reports fixed an issue where report tables with pagination enabled did not display referenced record tracking ids as clickable links in reference field columns reference field values now appear consistently when you page through report results content library / ssp content package deployment included unexpected components fixed an issue where deploying a content library package installed additional components beyond those you selected in the package deployment now includes only the components defined in the package you deploy playbooks were disabled after a content deployment error fixed an issue where playbooks were left disabled after a content library deployment failed partway through import, even when the failure was due to a missing application dependency playbooks affected by a failed deployment are no longer disabled unexpectedly classic playbook record trigger was missing after ssp import fixed an issue where importing a classic playbook from an ssp removed the record event trigger when associated applications were not included in the package record event triggers now remain on the playbook after ssp import when the trigger configuration is present in the package ssp export omitted required workflow dependencies fixed an issue where exporting an application ssp did not include required workflow dependencies referenced by the application, causing import to fail with a missing dependency error ssp export now includes required workflow dependencies so import completes successfully user management account page opened a different tenant than the one you selected fixed an issue where clicking a tenant on the account page sometimes opened a different tenant than the one you chose (for example, selecting uat and landing in sandbox), including a tenant you had opened earlier in the same session the tenant you select now opens consistently, and the toolbar shows the correct tenant name scim user list returned incomplete results fixed an issue where the scim users list endpoint returned no results or incomplete results when provisioning tools queried users through scim scim user queries now return the full set of provisioned users for the account disabled group status did not match user permissions fixed an issue where a disabled group in the admin ui did not accurately reflect whether members retained effective permissions, causing confusion about who could access tenant resources disabled group status now aligns with the permissions users actually have group assignment during user provisioning user provisioning now supports assigning groups during the provisioning workflow when the incoming payload includes valid group information, the user is created and associated with the specified group in a single operation if group data in the payload is invalid, the user is still created but is not assigned to a group role based access control home tab disappeared when switching tenants with enhanced rbac for accounts with enhanced rbac enabled, fixed an issue where the home tab was missing after you switched between tenants when your groups granted orchestrator access to different tenants the home tab now remains visible in each tenant you open enhanced rbac improvements fixed and improved multiple enhanced rbac behaviors from turbine 26 1, including permission aware navigation inconsistencies, issues in migrated and multi tenant environments, role validation and management workflows, and related permission enforcement edge cases hero ai / playbook builder max tokens field did not accept edits in hero ai action configuration fixed an issue where the max tokens field in the hero ai native action details dialog became unresponsive after you focused the prompt field first you can now edit max tokens without reloading the dialog html pasted into playbook template fields was rendered instead of saved as text fixed an issue where html pasted into template input fields in playbook configuration was rendered as formatted content instead of being stored as plain text pasted html is now treated as text in template inputs hero ai playbook builder applied changes beyond the requested scope fixed an issue where hero ai playbook builder modified multiple blocks or overwrote references when you asked for a single targeted change the builder now limits edits to the scope of your request and preserves unrelated configuration text2playbook changes were hard to review after generation fixed an issue where text2playbook did not clearly indicate which changes hero ai made versus manual edits, and visualization details referenced yaml structure that was difficult to act on during review change review now provides clearer, human readable feedback on what hero ai modified playbook generator used the wrong record identifier in record actions fixed an issue where hero ai playbook builder generated record action mappings using the internal record identifier instead of the tracking id field expected for record references generated record actions now reference the correct record identifier field text2playbook changed multiple elements when only one was requested fixed an issue where text2playbook updated multiple matching elements (for example, duplicate transformation actions in different blocks) when you asked to change only one targeted edit requests now change only the element you specify canvas removing a flow from a canvas solution left orphaned playbooks fixed an issue where removing a flow from a canvas solution did not delete the underlying flow playbook, leaving orphaned playbooks in the tenant removed flows are now deleted from the tenant when you remove them from the solution canvas solution did not open after a failed flow deletion fixed an issue where deleting a flow left the canvas solution in a broken state and the solution did not reopen when the underlying orchestration task deletion failed failed flow deletions no longer block access to the solution new canvas flow saved with the wrong enabled state fixed an issue where saving a new flow with a trigger in a canvas playbook sent the wrong disabled flag, so the flow did not match the enabled state you set in the editor new flows now save with the correct enabled or disabled state canvas solution saves were slow fixed an issue where saving a canvas solution from the solution builder took much longer than expected on some saves solution saves now complete in a reasonable time for typical solution sizes delete confirmation links opened the wrong playbook url fixed an issue where links in a delete confirmation dialog pointed to an incorrect playbook url and opened a 404 page confirmation links now open the correct playbook replacing parameters in playbook yaml added extra closing brackets fixed an issue where replacing an existing parameter in a playbook action duplicated closing curly brackets ( }} ) in the generated yaml, causing malformed configuration and run failures yaml generation now produces valid syntax when you swap parameters applet field names were truncated in the condition builder fixed an issue where applet fields in the condition builder displayed truncated object notation instead of the full field reference applet fields now show complete, usable references in conditions deleted transformation expressions remained referenced in advanced mode fixed an issue where deleting an expression in an advanced transformation action left a stale reference in the action configuration deleted expressions are now removed from references when you save the action transform data actions failed during agent restarts or high load fixed an issue where transform data native actions in playbook runs failed during agent restarts or high load transform data actions now retry and complete reliably during those conditions add expression inserted expressions missing the first character fixed an issue where add expression dropped the first character when inserting an expression (for example, inserting a b produced {{ b }} ) expressions now insert with the full reference you enter copy paste broke script task action input references fixed an issue where copying and pasting a script task that referenced action inputs replaced part of the reference with invalid placeholder text in the pasted script copied script content now retains valid action input references save stayed disabled after disabling a flow in older playbooks fixed an issue where disabling a flow in an older canvas playbook did not enable the save button, so the disabled state was not saved unless you made additional edits disabling a flow now marks the playbook as changed and enables save classic playbook conversion did not handle converging paths fixed an issue where converting a classic playbook with converging paths to canvas failed because multiple actions failed to connect to the same downstream action the conversion tool now duplicates actions after convergence so converted canvas playbooks build successfully converted playbooks missing application trigger button labels fixed an issue where converting a classic playbook with an application trigger left the trigger linked to the classic playbook, so the button label was missing on the converted canvas playbook application triggers now link to the converted playbook with the expected label converted playbooks gave every flow the same name fixed an issue where converting a classic playbook with multiple flows assigned the same flow name to every flow in the canvas playbook each converted flow now receives a distinct name classic playbook conversion failed when webhook url was not configured fixed an issue where the conversion tool attempted to convert classic playbooks that had a webhook trigger but no generated webhook url, producing an invalid canvas playbook playbooks without configured webhook urls are now excluded from conversion converted record event triggers used the wrong yaml property name fixed an issue where converted record event triggers wrote events instead of records in playbook yaml, preventing you from deleting and re adding the trigger converted triggers now use the correct property so trigger edits persist disabled classic playbooks were converted to canvas fixed an issue where the conversion tool converted disabled classic playbooks that were incomplete or not ready for migration disabled classic playbooks are now skipped during conversion orchestration webhook trigger logs showed no events after successful runs fixed an issue where webhook trigger logs reported no events even though the triggered playbook runs completed successfully webhook logs now show the events associated with successful runs orchestration alert email links opened the wrong run details page fixed an issue where links in orchestration alert emails opened an incorrect url and did not display run details in the alerts tab alert email links now open the correct run information asset search matched internal names instead of display titles fixed an issue where searching on the assets page matched connector asset internal names instead of the title shown in the list search now finds assets by the display title you see in the ui deleting an array item in an asset shifted password field values fixed an issue where deleting an item from an array in a connector asset left password field values on the wrong row and caused playbooks to read incorrect values from the asset array item deletion now reindexes all nested fields correctly native record actions showed a generic error message on failure fixed an issue where native record action failures (for example, patch or upsert) displayed only "an unexpected error occurred while executing the action " with no actionable detail in the run ui failed record actions now surface clearer error information in the playbook run view playbook runs slowed sharply or failed in nested loops with large items fixed an issue where playbook runs that used nested loop or repeat actions over large collections became progressively slower and could fail as iteration counts grew, especially when each iteration carried large item payloads loop runs now complete more reliably with less performance degradation playbook run job search api returned an internal server error fixed an issue where the playbook run job search api returned an internal server error for valid job search queries after upgrade to 26 1 2 the endpoint now returns matching jobs for the query you submit scheduled asset validation ran on an incorrect interval fixed an issue where the scheduled validation job for connector assets did not run on the intended hourly schedule after upgrade scheduled asset validation now runs every hour as configured connector installation failed when http proxy was configured fixed an issue on turbine on prem where connector installation failed because http proxy settings were not applied to the turbine api and swimlane tasks deployments, preventing package downloads connector installation now respects configured http proxy settings on those services file attachments failed when file api base url lacked a trailing slash fixed an issue on turbine on prem deployments behind certain load balancers or proxies where file operations failed because the file api base url was missing a trailing slash file api calls now use a correctly formed base url in helm deployments component import failed on self hosted environments fixed an issue on turbine on prem where importing certain components failed with item already exists errors even though the same package imported successfully in turbine cloud component import now completes on self hosted environments for affected packages playbook duplication showed an error and created an empty copy fixed an issue on turbine on prem where duplicating a canvas playbook displayed "system error occurred" and created an empty copy with no flows, actions, or triggers duplication now completes with the full playbook content copied known issues records concurrent record updates can conflict when the record has large text fields when multiple playbook runs use upsert or patch against the same record at the same time, and that record has large text fields, the updates can interfere with each other in rare cases a large text field value can be lost turbine 26 1 2 reduced how often this occurs but does not fully prevent concurrent update conflicts a longer term fix is planned for a future release playbooks required component inputs are not enforced at runtime when a component marks inputs as required in its input schema, playbook runs can still succeed if those values are not provided this is existing product behavior and is not changing in 26 2 0 because enforcing required inputs would be a breaking change for existing playbooks ctrl+c on canvas copies playbook content instead of selected text when you highlight text in a locked or disabled field on the canvas (for example, an asset name or action key), pressing ctrl+c or cmd+c copies playbook content from the canvas instead of only the text you selected workaround right click the selected text and choose copy that copies only the highlighted text turbine on prem tasks and api pods do not reach running state after deploying or upgrading turbine on prem 26 2 0, the swimlane tasks and swimlane api pods can remain stuck in init while waiting for the swimlane tenant service this can occur even when swimlane tenant pods report healthy liveness and readiness checks but the service is not yet accepting requests the issue is sporadic and does not affect every environment workaround restart the swimlane tenant deployment so the kubernetes service routes traffic to the new pods; swimlane tasks and swimlane api should then start normally upgrade to 26 2 0 can fail when completed kubernetes jobs remain upgrading turbine on prem to 26 2 0 can fail when completed jobs from a prior release remain in the cluster the upgrade may report that a job such as elasticsearch init is invalid because its configuration cannot be changed in place run the workaround below only when upgrading to 26 2 0; you do not need to repeat it on later upgrades workaround (before upgrade) in the swimlane namespace, run kubectl delete jobs field selector=status successful=1 , then proceed with the upgrade