Changelog

The functions of the Identity Management User Interface are now also fully available when using Keycloak as an Identity Provider. Additionally, the frontend has been expanded to include the tabs Organizations, Roles, and Permissions.

In these tabs, organizations, roles, and permissions can be

Furthermore, the users of these groups can be directly viewed and managed.

For organizations and roles (Keycloak as IdP only), role assignments can also be configured.

Further information can be found under Identitätsverwaltung.

The autoGrow option is now available for textarea.

It is now avoided that the selection is lost when editing the endpoint field (for HTTP proxies or IGUASU processors).

During the migration of BPC 4 settings, the inubit_dbTablePrefix is now set as a URL parameter at the httpProxy_referenceEndpoint.

This step is only performed when migrating from BPC 4 to BPC >= 5.0.7.

The error occurred when the primary key field on the monitor contained a field name that was partially capitalized. In this case, monitor actions could not be executed because the associated data could not be found.

fieldset now automatically insert scrollbars. Previously, scrollbars were automatically set only for container.

An incorrect value in settings of type linkedModules no longer prevents the picker from being opened.

Endpoints in the monitor, as well as HTTP proxy URLs, now support additional query parameters.

This can be useful, for example, to pass additional parameters to the INUBIT workflow via httpProxy_referenceEndpoint.

The INUBIT workflow package now correctly accepts BPC 5 calls.

Most workflows are now no longer delivered in a protected state.

The property names of the records in the payload of process actions and process starters with context retain their original format and are no longer converted to lowercase.

If Keycloak is configured as the identity provider, new and existing roles can now be assigned to and removed from users in the BPC frontend. When creating new roles, they are created as realm or client roles depending on the configuration. In user management, the necessary Keycloak roles must be taken into account.

The File Storage API can now create entries by passing a cloud reference, even if the referenced file does not yet exist. To do this, the query parameter check-file must be passed with the value false at the endpoint POST /cxf/bpc-core/file-storage/existing-file. This flag can also be provided in the Log Service when passing files as a cloud reference.

This can be useful if the file is uploaded to the cloud provider at a later time or asynchronously.

Further information can be found under POST in Log Service with Cloud Storage Reference and under File Storage API.

New window for conflict resolution when enlarging cells in the table layout, where decisions regarding the removal, moving, and shrinking of widgets can be made.

Analogous to BPC-9821, in addition to the BPC-Core, a Bad Request (400) error is now also returned in the REST endpoints of the corresponding modules if the HTTP body cannot be successfully deserialized.

The state of the action buttons on the replication jobs has been corrected. It could happen that changes to the replication jobs did not affect the availability of the functions.

In the cache configuration, services that no longer exist were being accessed, which leads to errors. These accesses have been removed.

It could happen that when calling modules, the navigation display was not updated correctly.

When exporting child data from the monitor, only the children of the parent entries that are visible according to the configured filters and views are exported.

The menu for monitor views no longer displays an empty placeholder when no view is selected.

The property names of the records in the payload of process actions and process starters with context retain their original format and are no longer converted to lowercase.

Previously, the OpenSearch internal field _id was always added to the monitor data. This is an unexpected extension of the actual document.

If this field is used in the monitor configuration, it should be switched to the correct primary key of the data. When using the Log Service, this can be found in https://docs.virtimo.net/de/bpc-docs/5.0/core/admin/modules/log_service.html#parentKeyFields. When using replication, in https://docs.virtimo.net/de/bpc-docs/5.0/core/admin/modules/replication.html#idColumns.

When starting Karaf, a Failed to process event error could occur under certain circumstances.

Update OpenSearch according to the update guide

During export, identical warnings and errors are now logged only once instead of for each exported entry.

In WebTSM, the interface texts have been revised to ensure consistent support for multiple languages.

Under certain circumstances, it could happen that when starting Karaf, the Backup Manager would not start due to an error.

The display of texts in the interface has been revised so that texts are fully visible.

If the index page was called by the browser when loading the BPC Web-GUI due to incorrect URLs for resources (images, scripts, stylesheets, …​), problems could occur during the verification of the login. This could only occur with OIDC-based identity providers.

A possible trigger can be, for example, a faulty theme.

The configuration org.apache.karaf.shell.cfg in Karaf has been revised.

Host-Key & Signatures: Switched to RSA 4096 and modern SHA-2 procedures (rsa-sha2-512/256); deactivated insecure NIST curves (ECDSA).

Protocol Hardening: Restricted to secure key exchange procedures (Curve25519), AES-GCM encryption, and Encrypt-then-MAC (EtM).

Important - If Karaf is not completely updated and only the configuration file is adopted, then the file etc/host.key must be deleted.

The payload of the form submit now contains the actionConfig again and thus also the ID of the process action.

Fix: During an update, the settings apiKeys (Core) and rendererUrl (Forms) are correctly migrated. apiKeys is converted from a map to an array, and in rendererUrl, "bpc-forms-renderer/index.html" is set as the value.

A problem has been resolved where larger amounts of data were only partially transferred in the HttpProxy.

After a BPC update to version 5.0.4 or 5.0.5, an error HTTP ERROR 500 javax.servlet.ServletException: javax servlet.ServletException: java lang.NoSuchMethodError: 'java lang.String de.virtimo.bpc.api.service.CoreBundleService.getUserFlowIdentityProviderLoginRedirectUrl(javax.servlet.http.HttpServletRequest)' could occur when accessing the BPC interface. This error has now been resolved.

The error occurred when a part of the interface (index.jsp) was not directly updated by the update and an OIDC/Karaf Identity Provider was used.

If an update to this version is not possible, the error can be resolved by deleting the directory <BPC-INSTALL-DIR>/karaf/data/pax-web/ and restarting the Apache Karaf.

Fix: Secure decoding prevents processor loading in monitor settings from failing

Radiogroups crashed the application on startup when a fixed value was used instead of data binding as the value. This issue has been resolved.

Error messages are correctly displayed again when starting processes. File downloads via action responses are working again. The loading of value lists for process starter comboboxes has been fixed. The success field of the return supports regular booleans (true/false) as well as strings ("true"/"false").

The GET endpoint of the Storage API can now also filter by value and customFields using the query and filter parameters. To filter by subfields, use dot-separated syntax: value.name:Alicia filters entries where the name field in value has the value Alicia.

Since storage indices do not contain timestamp fields, the parameters timezoneOffset and timezoneName have been removed from the GET endpoint.

For BPC Module Developers If you use the StorageService class and its getItems functions in your BPC modules, please note that these have been marked as deprecated and have been replaced with a variant without the timeZoneId parameter.

The developer documentation has been restructured: core:/dev/tutorials/index.adoc

The monitor’s caching mechanism (state in local storage) in interaction with the monitor settings has been precisely described.

Correction of the Replication documentation regarding the Replication_AdjustUpperDateLimitInSeconds setting, which is not limited to specific database management systems.

The documentation for monitor views has been updated. See monitor:/admin/process_views.adoc

With this update, the version of Apache Karaf is raised to 4.4.11.

Please update Karaf according to the update guide.

Incorrect behavior in the monitor regarding the data_filter setting has been fixed: If you used OR as the conjunction and did not set a base filter, all records were displayed despite filter rules.

With an AND conjunction, the base filter was negated and combined with the other filter rules using a logical AND. This was not correct; the base filter is now also not negated here and indicates which documents are generally visible. This means that a record is visible if it is matched by the base filter OR by all session-specific filter rules.

Attention: If you use monitor instances with data_filter in your BPC deployment, please verify whether they still function as intended.

More details about the data_filter setting can be found here: Restrict visibility of monitor data

For the monitor setting function_gridMultipleViewsConfig, a JSON schema has been provided. This offers guidance during configuration and helps to avoid incorrect configurations.

In TailSync, the database connection was not closed correctly. This leads to all connections in the connection pool becoming blocked over time.

Column headers and, in some cases, contents were unintentionally displayed centered. The background of the filters on checkbox columns was sometimes not displayed correctly.

Incorrect behavior in the monitor regarding the data_filter setting has been fixed: If you used OR as the conjunction and did not set a base filter, all records were displayed despite filter rules.

With an AND conjunction, the base filter was negated and combined with the other filter rules using a logical AND. This was not correct; the base filter is now also not negated here and indicates which documents are generally visible. This means that a record is visible if it is matched by the base filter OR by all session-specific filter rules.

Attention: If you use monitor instances with data_filter in your BPC deployment, please verify whether they still function as intended.

More details about the data_filter setting can be found here: Restrict visibility of monitor data

See Flow

The use of Flow for Forms has been described in detail. See Connection of IGUASU to the BPC

See core:/admin/operation/bpc-opensearch-tool.adoc

The documentation on multiple grid settings (function_gridMultipleViewsConfig) has been updated. See Configuration of multiple grids/views

See OpenSearch

Example of a dynamic filter value added with #user.loginName#. See Restrict visibility of monitor data

Data used for login with OIDC providers is stored encrypted in a cookie on the client. This prevents overload from a very large number of login requests.

For BPC module developers

If you have developed your own UserFlowIdentityProvider implementation, you need to adapt it. The function URI createAuthenticationRequestURI(String requestUrl) has been replaced with UserFlowLoginContext createAuthenticationRequestContext(String requestUrl) and now, in addition to the redirect URL for the OIDC login, also returns a cookie containing the login state. This cookie is set by the frontend during the login attempt.

For BPC administrators

Install the new modules bpc-be-core.jar and bpc-fe-core.war. A restart of the BPC/Karaf is necessary. If the BPC is still not available afterwards, please delete the directory karaf/data and restart Karaf again. (You may want to back up your log files beforehand.)

In process monitoring, the settings inubit_referenceEndpoint, inubit_proxyId, inutbit_baseUrl for downloading via an HTTP proxy have been consolidated and renamed to httpProxy_referenceEndpoint to clarify the possibility of downloading via a generic HTTP proxy connection. The setting inubit_dbTablePrefix has been removed, as it was only used for legacy scenarios.

Timeseries Management administrators who do not belong to the "bpcadmin" role and previously only had the "webtsm_admin" permission now need the "webtsm_edit" permission instead. A Timeseries Management administrator could also be assigned the "WEBTSM_ADMIN" role instead of the specific permission. This would also be more future-proof, for example, if in subsequent versions more fine-grained permissions are defined for the endpoints instead of the "webtsm_edit" permission.

The Data Management module no longer has a standalone backend component (bpc-be-vam.jar). This component must be removed during the update.

To perform the update, first remove bpc-be-vam.jar and then deploy the updated bpc-fe-vam.war.

The Data Management module has now been integrated into the regular release cycle.

The setting baseUrl is omitted, as the value (cxf) is a static part of the API URL and does not change.

Do not confuse this with bpcBaseUrl. This setting remains unchanged.

The format for index imports has been changed to a ZIP-based format. This is the same format that our opensearch-tool also uses to import and export indices. In the administration interface under Core Services → Indices, such index export files can be downloaded.

Note: If you use the index import functionality for an installation with a preconfigured BPC, the export file must be recreated.

See also Installation with preconfigured BPC.

Files in File Storage can now be created not only via direct upload but also through a reference to a file in the cloud storage provider. For this purpose, a new endpoint is offered in the File Storage API. In the Log Service, files (for fields of type file-storage) can also be transmitted as a file reference to the cloud storage provider.

Additionally, when uploading and downloading with all supported cloud providers, the content type is set.

File storage backend connections can now be configured so that downloads are routed through the BPC. See core:/admin/modules/backend_connections.adoc#file_storage. Downloads via the file storage service are now documented in the audit log.

The File-Storage Log Service integration has been adapted for the Process Monitor so that it can offer these files for download and viewing: For fields of type file-storage, the Log Service now creates two additional fields with the suffix _filename and _contentType when logging in OpenSearch and relational databases.

Files from the File Storage can now also be accessed in the Process Monitor.

The format for index imports has been changed to a ZIP-based format. This is the same format that our opensearch-tool also uses to import and export indices. In the administration interface under Core Services → Indices, such index export files can be downloaded.

Note: If you use the index import functionality for an installation with a preconfigured BPC, the export file must be recreated.

See also Installation with preconfigured BPC.

Via the WebTSM setting module_apis, additional HTTP headers can be defined, which are then set for every WebTSM API call. To do this, include the field additionalHttpHeaders with the HTTP headers to be set (see example). Variables, such as the one below ${config.de.virtimo.bpc.core.karaf.uuid}, can be set in the same way as with the BPC HTTP Proxy.

The configuration of file attachments and downloads in the BPC Monitor has been restructured:

Further information can be found here: Configuration of the monitor columns and File attachments.

Attention: Monitor instances had provided the setting column_mimetype to specify the column name that refers to the content type of the files. However, in the backend, the column name contentType was used inconsistently. Therefore, during the migration of the instances, the column_mimetype setting was ignored. If this was used, please check the corresponding monitor instance and adjust it if necessary.

Sharing links for monitor processes have been shortened. The link now only contains information about the active view, the filters, and the sorter.

Form versions and forms can now be deleted in the Forms Repository.

disabled and readOnly are available for table columns. disabled, readOnly, and required are now only inherited for configuration in group components. Attributes are not overwritten in this process.

You can now configure the initial sort order for groups in the monitor settings. See setting column_groupDirection

Data used for login with OIDC providers is stored encrypted in a cookie on the client. This prevents overload from a very large number of login requests.

For BPC module developers

If you have developed your own UserFlowIdentityProvider implementation, you need to adapt it. The function URI createAuthenticationRequestURI(String requestUrl) has been replaced with UserFlowLoginContext createAuthenticationRequestContext(String requestUrl) and now, in addition to the redirect URL for the OIDC login, also returns a cookie containing the login state. This cookie is set by the frontend during the login attempt.

For BPC administrators

Install the new modules bpc-be-core.jar and bpc-fe-core.war. A restart of the BPC/Karaf is necessary. If the BPC is still not available afterwards, please delete the directory karaf/data and restart Karaf again. (You may want to back up your log files beforehand.)

Only the WebTSM backend communicates directly with Hakom. The WebTSM frontend now only communicates with the WebTSM backend. This means that a backend connection of the type "HTTP Proxy" is no longer necessary and the setting "tsm_backendConnectionId" has been removed from the WebTSM Viewer components.

If "useRendererValues" is true during the process status change, the fallback values "" and "*" are no longer displayed in the status combo box.

In process monitoring, the settings inubit_referenceEndpoint, inubit_proxyId, inutbit_baseUrl for downloading via an HTTP proxy have been consolidated and renamed to httpProxy_referenceEndpoint to clarify the possibility of downloading via a generic HTTP proxy connection. The setting inubit_dbTablePrefix has been removed, as it was only used for legacy scenarios.

Multilingualism has been improved. The languages cs, da, de, en, es, fi, fr, it, js, ko, nl, pt, ru according to the ISO 639-1 language codes are now available. If a language is not available or incomplete, English will be used as a fallback.

Please remove the values with the key "authorization" in the Timeseries Management JSON setting under General → module_apis. This value contains the username and password in plain text (Base64 encoded) and is no longer needed.

Timeseries Management administrators who do not belong to the "bpcadmin" role and previously only had the "webtsm_admin" permission now need the "webtsm_edit" permission instead. A Timeseries Management administrator could also be assigned the "WEBTSM_ADMIN" role instead of the specific permission. This would also be more future-proof, for example, if in subsequent versions more fine-grained permissions are defined for the endpoints instead of the "webtsm_edit" permission.

Through this change, tables in the GUI can be tested more effectively in an automated manner.

The setting baseUrl is omitted, as the value (cxf) is a static part of the API URL and does not change.

Do not confuse this with bpcBaseUrl. This setting remains unchanged.

Duplicate setting groups are merged into one group

If multiple columns are configured in a monitor instance that use the same dataIndex but are displayed differently by custom renderers, this is now correctly taken into account during export. Additionally, during export, mappings for empty values and fallback values in custom renderers are considered (entries "!" and "*").

The error in the schema generation of some DTO classes has been fixed.

When exporting child data from a monitor, the data_filter setting is applied to the corresponding parent entries.

Filtering on the column is now possible again. An additional option has also been added to set a different template for the filter. This may be necessary if the column template accesses fields other than those specified in dataIndex.

See also Configuration of the monitor columns

targetModule is automatically converted to a string in the viewConfiguration

Securing BpcCommon.Api.getTranslation: If a translation incorrectly returns an object (e.g., for config), it now falls back to the key string to avoid [object Object].

When changing the process status, it is now also possible to use "preselectState": "_first" if "useDistinctValues" is false and custom states are used. See also Process Status Change Configuration

The spacing issues of checkboxes within a checkbox group have been fixed.

Previously, only columns in the child data were exported for which there are also columns in the parent monitor that use a dataIndex with the same name.

An export in the Process Monitor now works even if instance_data_view_limit or monitor_data_view_limit is less than 2500.

The documentation on binding in Forms has been expanded.

See:

See Status Change - Syntax

Update OpenSearch according to the update guide

With this update, the version of Apache Karaf is raised to 4.4.10. In this context, it is necessary to update the following modules: bpc-be-core, bpc-be-analysis, bpc-be-forms, and bpc-be-monitor.

Please update Karaf according to the update guide.

Adjustment of the logging configuration

The configuration file etc/org.ops4j.pax.logging.cfg was adjusted as part of a Log4J update. Important: Do not simply copy this file unchanged from an existing (old) Karaf installation. If you have made individual adjustments, please transfer them manually to the new configuration file.

Note for BPC module developers

If your BPC backend module uses the dependency jackson-annotations and has so far used the variable ${jackson.version} (to reference the same version as in the BPC core modules), this needs to be adjusted. Instead, use ${jackson.annotations.version}.

The Data Management module no longer has a standalone backend component (bpc-be-vam.jar). This component must be removed during the update.

To perform the update, first remove bpc-be-vam.jar and then deploy the updated bpc-fe-vam.war.

The Data Management module has now been integrated into the regular release cycle.

The URLs for Flow endpoints can be copied directly from the Flow Manager to the clipboard and reused, for example, for form submission URLs.

The prometheus-exporter plugin is now included by default with OpenSearch.

The BPC API now also supports authentication using access tokens issued by the configured OIDC provider. Authentication via access tokens must be explicitly enabled in the identity provider configuration. In addition to validating the signature and validity of the access token, the OIDC provider’s introspection endpoint can also be used, which additionally checks whether the session has been terminated (token revocation). Further information can be found here: Access tokens and Keycloak as an identity provider.

In the frontend, the global event beforeShowModule is triggered before switching to a new view (module or module component). The following parameters are passed: the module ID of the current view, the module ID of the new view to be displayed, and additional options. If an event listener returns false, the switch to the new view is canceled. If this behavior is implemented, the user should also receive feedback explaining why the switch was canceled.

Starting with BPC 4.2.13 and 4.1.25, process actions were no longer displayed in child grids. This issue has been resolved, and process starters are no longer incorrectly displayed in the child grid.

The behavior introduced in BPC 4.2.16 and 5.0.2, whereby the connection is explicitly closed in the backend connection, sometimes led to issues where responses from the server were not transmitted at all or were transmitted incompletely.

The BPC Breadcrumb Plugin is displayed again

The BPC File Storage Service now supports Azure Blob Storage and Google Cloud Storage in addition to AWS S3.

To use the file storage functionality, it is necessary to install the appropriate module for the respective cloud provider, see File Storage Modules. (This is also required if file storage was previously used with AWS S3.) If file storage is already being used with AWS S3, the AWS S3 File Storage Module must also be installed accordingly.

A security check has been introduced that verifies HTTPS is used for the bpcBaseUrl setting.

It is now possible to store additional headers on the HTTP proxy, which will be sent to the server.

Analogous to the import of module settings, we now offer the possibility to import preconfigured OpenSearch indices, see Installation with preconfigured BPC.

A command line tool is available for listing, deleting, as well as exporting/importing indices. It is called bpc-opensearch-tool.jar and can be downloaded from the file server (https://files.virtimo.net/bpc/). You can access the help by running java -jar bpc-opensearch-tool.jar --help.

When creating a Logservice instance, the External Reference field is now included in the default configuration. In addition, the field types object and flat_object have been added.

A security check has been introduced to ensure that BPC is only accessible via HTTPS.

The delivered configuration file de.virtimo.bpc.core.security.checks.cfg contains corrected and updated default settings adapted to current checks.

Tooltips with URL links can be accessed with the cursor without them disappearing.

Fix for a potential DoS attack that leads to high memory consumption when there are a large number of parallel log-in requests.

The configured limit (max. number of records) for XLSX is now taken into account again during data export in the monitor. For CSV and JSON, a limit can now also be configured in the export mask. There is only an upper limit for XLSX, not for other file formats.
See also monitor:admin/configuration_monitor_export.adoc
Export of processes

When exporting monitor data with configured custom renderers, the following errors were fixed: - Custom renderers were only applied to data of type String. Now, custom renderers are also applied to the types Integer, Long, and Boolean during export. - For a column, you can set the value true as a custom renderer. In this case, a custom renderer with the same name as the dataIndex is used. If this does not exist, the custom renderer "defaultRenderer" is used. During export, only "defaultRenderer" was used in this case. This has been fixed.

For GUI modules that offer types, components with types can once again be created. Translations for the designation of a type no longer require a "CORE_" prefix. The schema for the translation keys is "<MODULID>_INSTANCE_TYPE_<VALUE>".

Fixing an exception during configuration update when the Security Cockpit has not yet been opened.

Security check measurements now record a timestamp again.

The HttpProxySecurityCheck is executed only once per instance.

The default instances in the Analysis module are now correctly initialized, so that the data modifier is immediately available.

A faulty plugin configuration is now ignored during migration and logged as a warning.

The "Users" section in the documentation page "BPC Administration" is described in more detail. See BPC Administration

Extended description of the process action configuration. See Configuration of process actions

Detailed documentation of the process status change configuration. See Process status change

See monitor:admin/configuration_monitor_export.adoc

The documentation of formatters in columns and formatters in export has been improved.

The lists of the respective module settings have been revised or are now generated automatically. This ensures that the identifiers and description texts are identical to those in the GUI. See Configuration of Data Analysis Core Services settings Configuration of the Process Dashboard Configuration of Forms Configuration of Process Monitoring

See Chart Editor

Check and, if necessary, adjust the HTTP proxy configurations. Up to now, a '/' was always appended to the configured connection URL. This is no longer the case. If your HTTP proxy target requires this and the URL is currently configured without a '/' at the end, please adjust it accordingly.

Two new plugins have been added to link to internal or external content. See also External Link and Internal Link

In the frontend, it is now possible to manually start a tail sync for replication jobs. Additionally, a manual tail sync can also be started via the REST API; see Replication API.

BPC now supports PKCE (Proof Key for Code Exchange) for OpenID Connect and Keycloak Identity Provider to enable a hardened login flow. To activate this, you need to set the PKCE method in the Identity Provider backend connection and, if necessary, configure this in your Keycloak client.

Tail sync runs can now be logged. When logging is enabled, entries are written to the index bpc-tailsync-log. Additionally, the monitor Tail Sync Log Monitor is created. You can find more information about this under Replication.

The marking of settings that have values differing from the default has been changed.

In the form for sending notifications, the role bpcuser is preselected as the recipient group. It is no longer possible to send a notification without any recipient group.

When reindexing, the old index is deleted instead of being closed as before. The user is offered the option to create a backup before reindexing.

The values of the different options in combo boxes are now no longer generally supplemented with a  . This makes automated testing of the interfaces easier.

Code that was marked as deprecated has been removed from BPC.

The driver provided via pax-jdbc-mssql is no longer delivered.

Please ensure that the MSSQL driver used has at least one of the following versions: 10.2.4, 11.2.4, 12.2.1, 12.6.5, 12.8.2, 12.10.2, 13.2.1

If an older driver is being used, it should be replaced. Remove the old driver by deleting it from the deploy directory or by uninstalling pax-jdbc-mssql or the driver itself. Then install a current driver (without the pax-jdbc-mssql feature). See also: Databases

A bug was fixed where an incorrect process starter was displayed when exactly two process starters were configured, one of which is hidden.

Check and, if necessary, adjust the HTTP proxy configurations. Up to now, a '/' was always appended to the configured connection URL. This is no longer the case. If your HTTP proxy target requires this and the URL is currently configured without a '/' at the end, please adjust it accordingly.

If an error occurs during the migration of an index, the newly created index for the migration will be removed again in order to avoid subsequent errors or to revert to a stable state.

Buttons that discard local changes now have a better text description.

Changes to the valid value range have a direct impact on the validation of the default value.

In the configuration file, the value for deleteCheckResultsOlderThan must be changed from 10 days to 10 days ago.

By default, the accessibility of Karaf via SSH connections is now restricted to the host localhost. To enable accessibility via external connections, use the central configuration file. See Karaf SSH Access.

Connections to OpenSearch are now configured exclusively via de.virtimo.bpc.core.opensearch.hosts. When updating from an older version, you must add the entry de.virtimo.bpc.core.opensearch.hosts to the configuration file karaf/etc/de.virtimo.bpc.core.cfg. Additionally, it is recommended to supplement the central configuration file with the lines

or

(as shown here with the example https://localhost:9200). You can use the files from the download file server as a reference.

The use of Java 21 is now necessary.

These metrics have been renamed:

INUBIT is no longer supported as an Identity Provider.

For the Identity Provider (Backend Connections), a large part was previously configured via a JSON setting. This has been split into individual settings. Existing configurations are migrated automatically. If older deployment exports of Identity Provider components are used, it is best to recreate them.

See also: Create, configure and use an identity provider, Database as identity provider, Keycloak as an identity provider, Keycloak and dynamic redirect URIs, Deployment constraints

The endpoint /cxf/bpc-core/status/server/{serverUUID} of the status API, which queried the status of a remote BPC, has been removed.

The query string parameters gridId, gridExtId, multiRecords, buttonId and the form parameter tablePrefix have been removed from the process action requests. To send additional context or configuration information, "process action metadata" can be configured in the monitor settings.

In process starters, the XML that was sent has been replaced by JSON. As a result, portletArchiveName, operation, mandant, gridID, key, bpcModule, bpcModuleInstanceId and custom have been removed. key has been uniformly renamed to id. Additional context or configuration information can be reliably sent via the "metadata" configuration in the process starter configuration in the monitor settings. The process parameters are bundled under config.parameters in the request payload. File uploads contain filename, type and data. Grids contain a list of their records.

In Change State, tablePrefix, mandant and changeStatusBox_<column name> have been removed. command has been uniformly renamed to type and receives the value "statusChange". columnsstring has been renamed to column. newStatusCombo_<column name> has been renamed to newStatus. commentfield has been renamed to comment. childStatus is no longer set to "Info" by default in the Change State configuration and, if configured, is migrated to metadata. Additional information can be passed in the "metadata" object in the Change State configuration.

See also Migration from BPC 4.* to BPC 5.0 Configuration of process actions Process Starter

The unused settings inubit_aperakEndPoint, inubit_dbGridId and inubit_pmMandant have been removed. The settings inubit_actionEndpoint, inubit_changeStateEndpoint and inubit_VpsEndpoint have been replaced by actionEndpointProcessor, changeStateEndpointProcessor and vpsEndpointProcessor.

HTTP proxy backend connections and flow connections now always filter the session cookie so that a recipient cannot make calls in the BPC in the context of the user. The setting filterSessionCookie is therefore no longer required. Instead, with the new setting sendSessionId, it is possible to send the user’s session ID, which can be verified at the endpoint GET /cxf/bpc-core/authentication/session/{sessionid}. When the injectUserSessionJWT setting is enabled, a self-created JWT is no longer sent, but rather the signed ID token from the OpenID Connect provider is sent. You can find more details in Backend Connections - HTTP-Proxy.

Http calls via an Http-Proxy or Flow connection filter out the BPC Api-Key header (X-APIKey). Attention: This could affect existing BPC configurations, for example if an INUBIT process is triggered that in turn makes calls to the BPC API using the provided API key. (In this case, it would be better to store a fixed BPC API key in the INUBIT process.)

Users of the Log Service API must replace all occurrences of childs with children in their POST data. During a transition period, childs can still be used for incoming data. This backward compatibility will be removed in a future release. It should also be noted that the responses from the Log Service endpoints now contain children instead of childs.

The OpenSearch configuration directory can be configured via OPENSEARCH_PATH_CONF. This allows you to outsource the configuration directory from the OpenSearch directory. As a result, you will no longer overwrite it during an OpenSearch update. See also Outsourcing OpenSearch Configuration.

In the bundle installation file, BPC is now delivered with a configuration directory outsourced to INSTALLATION_DIRECTORY/opensearch_config.

It is recommended to set OPENSEARCH_PATH_CONF in the bpc.env.

A new action type "bulkAction" has been added to enable actions to be performed for all records. See Bulk Actions

If IGUASU is correctly connected via the Flow module, available processors can easily be selected through the configuration interface of the monitor actions.

A configuration interface for identity providers is now available.

With this update, it is necessary to update the Karaf.

The BPC provides OpenApi-compliant specification files for our APIs. These can be found under Downloads and can alternatively also be accessed dynamically via the BPC, provided this option is enabled. You can find more information in the section BPC-API.

A Security Cockpit is now offered in the administration area, providing an overview of application security.

For details, see Security cockpit

A new API endpoint /cxf/bpc-core/collaboration/online-users provides the ability to retrieve all logged-in users. (See Collaboration API.) A new plugin "List Online Users" makes it possible to display all logged-in users in the frontend. (See Online User List))

A chat plugin has been added. The chat is stateless, so messages are lost when you log out, reload, or close the page. New messages are only received while you are logged in. See also Chat

You can use the Log Service API to be redirected directly to the Log Service configuration or to connected monitors. There are two new LogService endpoints that redirect the user to the corresponding BPC pages when called:

See also API documentation: Log Service API

It is now possible to specify references to external resources when writing audit information. For example, it is possible to refer to the IGUASU instance that created the entry.

Replication now also supports, as an alternative to existing database tables/views, the direct entry of an SQL query.

See also sourceCommonTableExpressionQuery in Replication

A JSON schema can now be requested via the Log Service API, which describes the payload for the Log Service. This can be used to perform a validation in advance.

A new setting has been added that allows you to select the endpoints or processors of Flow & Backend Connections. The selection generates a BPC URL following the pattern bpc://<flow/backendconnection>/<instanceId>/<endpointOrProcessor>. There are separate settings for process actions, process starters, and change state.

Messages can be sent to individual users. Messages can be time-limited. Messages can set a topic. It is possible to delete all messages of a single topic. The status of whether a message has been read is now persisted in the backend, instead of being stored in the browser’s local storage.

See also Notifications to users and Notification API for changes to the API.

The backend connections are being extended by a 'file_storage' type to configure connections to AWS S3 and S3-compatible file storage systems.

A file storage API is provided that enables uploading, downloading, and managing files. These are stored on a file storage configured via a file storage backend connection.

See File-Storage API.

Under Settings→Overview→File Storage, a UI is provided to display, download, and delete files from the file storage. Additionally, files can be uploaded to the file storage.

File Storage Integration in the Log Service

Further information can be found in File Storage in the Log Service

In the frontend, it is possible to redirect to a BPC URL after the login process using the query parameter redirectPostAuth. This works for all OIDC identity providers.

Results from the validation are now stored in the state. This means they can also be used for bindings, for example. See also validate

A new setting has been added that allows you to select the endpoints or processors of Flow & Backend Connections. The selection generates a BPC URL following the schema bpc://<flow/backendconnection>/<instanceId>/<endpointOrProcessor>. There are separate settings for process actions, process starters, and change state. The URL is always uniformly included in the payload during the POST request and passed to the httpProxy endpoint of the monitor. The monitor endpoint uses the Core Action Service to further process the data and ultimately returns the response.

BPC API requests that take place via a web browser now redirect to Keycloak (or another OIDC UserFlowIdentityProvider) if authentication is missing. After logging in, a redirect to the original API endpoint takes place.

API keys are now masked in the frontend, just like passwords. When viewing existing API keys in settings, is now displayed instead of the API key.

When a new API key is created, it is displayed once.

See System monitoring (metrics for Prometheus)

If the available disk space falls below defined thresholds, shards are redistributed to other nodes. It can also happen that indices are set to read-only to prevent the disk from filling up.

The value of cluster.routing.allocation.disk.threshold_enabled is now true. This corresponds to the OpenSearch default value.

The OpenSearch configuration directory can be configured via OPENSEARCH_PATH_CONF. This allows you to outsource the configuration directory from the OpenSearch directory. As a result, you will no longer overwrite it during an OpenSearch update. See also Outsourcing OpenSearch Configuration.

In the bundle installation file, BPC is now delivered with a configuration directory outsourced to INSTALLATION_DIRECTORY/opensearch_config.

It is recommended to set OPENSEARCH_PATH_CONF in the bpc.env.

Connections to OpenSearch are now configured exclusively via de.virtimo.bpc.core.opensearch.hosts. When updating from an older version, you must add the entry de.virtimo.bpc.core.opensearch.hosts to the configuration file karaf/etc/de.virtimo.bpc.core.cfg. Additionally, it is recommended to supplement the central configuration file with the lines

or

(as shown here with the example https://localhost:9200). You can use the files from the download file server as a reference.

The setting Common Table Expression Query in replication can now be edited in an editor that supports SQL syntax.

These metrics have been renamed:

The standard mapping (Mapping of organizations, roles and rights) for Identity Provider has been adjusted so that every user (via the standard role bpcuser) receives the right to use all standard modules of the BPC. Additionally, the right to view the list of logged-in users is granted.

This change does not affect existing configurations.

For the Identity Provider (Backend Connections), a large part was previously configured via a JSON setting. This has been split into individual settings. Existing configurations are migrated automatically. If older deployment exports of Identity Provider components are used, it is best to recreate them.

See also: Create, configure and use an identity provider, Database as identity provider, Keycloak as an identity provider, Keycloak and dynamic redirect URIs, Deployment constraints

The setting of the ID columns in the log service configuration interface is now done via combo boxes. The configuration can only be saved if at least one ID column is defined. Additionally, it is checked whether all ID columns are included in the field configuration. If this is not the case, the form is marked as invalid and saving is prevented.

The HTTP proxies configured in BPC now also forward OPTIONS and PATCH requests.

For notifications, the type 'link' no longer exists; instead, all notifications can now specify links. The field linkData is used for this purpose (previously typeSpecificData). For references to external websites, the url field is still used; for BPC-internal navigation, the hash component of the BPC URL is now specified in the navigation field. You can read more in Notification Types and in Notification API.

In the notification interface (BPC → Configuration → Overview → Notifications), a link can now also be configured (see Notifications to Users).

The Submit has been revised. Some changes are now required here. See Migration from BPC 4.* to BPC 5.0. The structure and possible actions of the messages have been adjusted.

The endpoint /cxf/bpc-core/status/server/{serverUUID} of the status API, which queried the status of a remote BPC, has been removed.

The plugin can now configure a process that is displayed immediately upon triggering. Additionally, the text and icon can now be freely configured. See also Process Starter Plugin

The option for grouping settings is now initially enabled. It is recommended to use this view, as the settings are displayed together in a common context.

There are now multilingual labels on groups of settings.

Log services can now be (de)activated individually. The term maintenance mode was misleading in this context.

The query string parameters gridId, gridExtId, multiRecords, buttonId and the form parameter tablePrefix have been removed from the process action requests. To send additional context or configuration information, "process action metadata" can be configured in the monitor settings.

In process starters, the XML that was sent has been replaced by JSON. As a result, portletArchiveName, operation, mandant, gridID, key, bpcModule, bpcModuleInstanceId and custom have been removed. key has been uniformly renamed to id. Additional context or configuration information can be reliably sent via the "metadata" configuration in the process starter configuration in the monitor settings. The process parameters are bundled under config.parameters in the request payload. File uploads contain filename, type and data. Grids contain a list of their records.

In Change State, tablePrefix, mandant and changeStatusBox_<column name> have been removed. command has been uniformly renamed to type and receives the value "statusChange". columnsstring has been renamed to column. newStatusCombo_<column name> has been renamed to newStatus. commentfield has been renamed to comment. childStatus is no longer set to "Info" by default in the Change State configuration and, if configured, is migrated to metadata. Additional information can be passed in the "metadata" object in the Change State configuration.

See also Migration from BPC 4.* to BPC 5.0 Configuration of process actions Process Starter

The unused settings inubit_aperakEndPoint, inubit_dbGridId and inubit_pmMandant have been removed. The settings inubit_actionEndpoint, inubit_changeStateEndpoint and inubit_VpsEndpoint have been replaced by actionEndpointProcessor, changeStateEndpointProcessor and vpsEndpointProcessor.

An invalid JSON payload in an HTTP request body is now responded to by API endpoints with the error 400 Bad Request.

The modelVersion is now also displayed when the caller does not have admin rights.

For the Forms module, a migration is required between BPC 4.x and BPC 5. Information about this can be found here.

For boolean values in the settings tables, a combo box is no longer displayed; instead, only the checkbox with the truth value is shown. The value can be changed by clicking, pressing the space bar, or pressing Enter.

The appearance of BPC has been updated.

Buttons can now use the classes arrow-spacing-sm (small) and arrow-spacing-xs (extra small) to reduce the spacing of the arrow.

For BPC module developers:

If a feature/feature.xml is used in the module sources, please remove the following resource block from the pom.xml:

This entry defined the feature.xml as a resource and embedded it into the JAR file. This is not necessary.

However, the maven-resources-plugin still needs to be used to replace placeholders in the feature.xml.

To do this, add the following plugin call:

For the migration of installations that were already using Elasticsearch, it may be necessary to perform an additional migration step when switching to OpenSearch 3. A corresponding tool is provided for this purpose. For further details, see Migration from BPC 4.* to BPC 5.0

Index cleanups are performed 60 seconds after BPC starts, and then run at the configured interval.

It is now possible to use the value 0 in the valueMap of a custom renderer and to see it as a label in the monitor. Previously, values of 0 were not displayed.

The symbol for accessing notifications and displaying the number of unread notifications is now a plugin. Therefore, like other plugins, it can now be placed in various locations in the interface.

For notifications, there is now an additional option to send the message only to logged-in users. The filters by roles and organizations remain active. See also Notifications. For notifications with the recipient group "Users," the recipients are no longer displayed.

Users of the Log Service API must replace all occurrences of childs with children in their POST data. During a transition period, childs can still be used for incoming data. This backward compatibility will be removed in a future release. It should also be noted that the responses from the Log Service endpoints now contain children instead of childs.

The unused setting identityProvider_ussUrl has been removed.

A JSON schema validation of the data to be written can be activated on each Log Service component. The default setting is 'Off'. The JSON schema is automatically generated based on the configured Keys and Fields and can be retrieved if necessary via the Log Service endpoint.

When handling events, Map should be used instead of Dictionary.

Any path specifications are ignored to prevent path traversal attacks.

By default, the accessibility of Karaf via SSH connections is now restricted to the host localhost. To enable accessibility via external connections, use the central configuration file. See Karaf SSH Access.

The logging configuration has been changed so that session tokens are masked. For new installations of Karaf, this happens automatically. For existing installations, please adjust the configuration file [KARAF]/etc/org.ops4j.pax.logging.cfg. Replace the old line

with the following lines

Since with Keycloak and OIDC the session tokens are UUIDs that are also used elsewhere, we log the first and last four characters here.

HTTP proxy backend connections and flow connections now always filter the session cookie so that a recipient cannot make calls in the BPC in the context of the user. The setting filterSessionCookie is therefore no longer required. Instead, with the new setting sendSessionId, it is possible to send the user’s session ID, which can be verified at the endpoint GET /cxf/bpc-core/authentication/session/{sessionid}. When the injectUserSessionJWT setting is enabled, a self-created JWT is no longer sent, but rather the signed ID token from the OpenID Connect provider is sent. You can find more details in Backend Connections - HTTP-Proxy.

Http calls via an Http-Proxy or Flow connection filter out the BPC Api-Key header (X-APIKey). Attention: This could affect existing BPC configurations, for example if an INUBIT process is triggered that in turn makes calls to the BPC API using the provided API key. (In this case, it would be better to store a fixed BPC API key in the INUBIT process.)

For new backend connections of the type HTTP-Proxy and Flow, the option Filter BPC Session is enabled by default.

If you use the file custom.java.security, you should adjust the property jdk.tls.disabledAlgorithms to make the TLS configuration stricter and therefore more secure.

Instructions for migrating old indices during the migration to OpenSearch 3. See also Migration from BPC 4.* to BPC 5.0

The use of Java 21 is now necessary.

If BPC was initially installed with an OpenSearch version instead of an Elasticsearch version, then the [bpc]/opensearch_data directory can be adopted directly.

If not, indices must first be migrated.

See also Migration from BPC 4.* to BPC 5.0

With this update, the Karaf version is upgraded to 4.4.8. It is necessary to update the modules bpc-be-core, bpc-be-analysis, bpc-be-forms, and bpc-be-monitor.

For BPC module developers

With the updated Karaf, we are delivering CXF 3.6.8 instead of 3.6.7, which provides Jackson in version 2.19.2.

Please update the CXF version (3.6.8) and Jackson version (2.19.2) in your pom.xml. Check whether you have an Import-Package statement for this version. This must be added or adjusted if you use Jackson features such as the ObjectMapper class.

The core is now no longer transpiled into old ECMAScript and no polyfills are added anymore. This enables support for newer language features. This means that ExtJS packages should also have the following content in their packages/local/PACKAGENAME/package.json:

INUBIT is no longer supported as an Identity Provider.

See OpenAPI specification for process actions, process starters, process status changes and multiple actions

The renderer of the Forms module is now no longer included in bpc-fe-forms, but is instead provided separately as bpc-forms-renderer. This makes it possible to update the renderer independently. When a BPC update occurs, the URL of the renderer in the Forms module is automatically changed to bpc-forms-renderer/index.html.

The renderer is now available via Downloads.

The detail view element type "taskPanel" has been removed along with the associated settings function_interactionColumn and function_interactionFormId.

For a transitional period, the use of childs instead of children is still permitted. This does not affect the data structure when reading data.