Verba Release Notes

New REST API endpoints for server configuration and credential management: - Change credentials such as usernames, passwords, secrets, and connection strings for users, storage targets, import sources, AD synchronization profiles - Import/export server configurations or configuration profiles - Change server configuration or configuration profile values - Execute server configuration tasks - learn more

Migrated to the latest (3.1.1) OpenSSL in C++ services.

OpenSSL library is updated to the latest version 3.2.0.

Enhancement to the Cisco Call Direction recognition. The recorder service now only considers an address as internal if: - CUCM signals that the address is internal, or - the address is set up as a recorded Extension, or - the address is observed by the JTAPI Service, or - the address matches the Internal Domain Numbers Pattern.

New alerts added to the Microsoft Teams Bot Service: - Call timed out: this is an extension of RI-030767, when the service is closing a stuck call after a configurable timeout (45 minutes by default), it sends an alert. - Could not authenticate: if the service is failing to request a new OAuth token for the outbound API requests, or if the service is not able to validate the inbound requests from Teams e.g: the service is not able to retrieve the OpenID Connect keys - Recorder overloaded: the service now sends an alert if any of the recorder connections are overloaded. The alert contains the current performance metrics shared by the Verba Unified Call Recorder Service - Unexpected call termination: the service sends an alert if the call termination is not normal (200) - The maintenance thread now checks the public certificate expiration and private key accessibility. If the certificate expires within 30 days the service sends an alert.

Support for the Microsoft Teams Large Gallery View displaying mode added

Added new selective recording mode to the Microsoft Teams Bot service. When the setting is selected in the Extension configuration, the bot service automatically starts controlled recording if an external participant joins the call. This recording mode applies to call scenarios such as PSTN, federated P2P calls, or meetings with at least one external participant.

The Microsoft Teams Bot Service now sends the ongoing sessions list to the Unified Call Recorder Service every 5 minutes.

Selective recording rules can now be configured for Microsoft Teams integrations.

The Microsoft Teams Bot service now supports the 'Record Calls Answered by 3rd Party' extension configuration.

Added TLS 1.3 support for communication between .NET and other VFC Services. Also updated the .NET Framework version to 4.8.

Microsoft Teams IM Export API enhancements: - Recently introduced system events in the Export API that cause duplicated event messages in the chat are no longer processed. - In case of API connection down alerts, more information is provided in the alert about the cause. - In case of unexpected API responses, VFC only attempts one retry in the same polling cycle to avoid excessive API usage.

Added audio and video clip content support for Microsoft Teams IM recording.

The Avaya DMCC Service no longer processes the single-step conference events generated by other recorder products.

IPTrade TPO: Implemented the parsing of additional metadata: conference participation, device ID, and turret ID.

Improved the performance of call segmentation for large amounts of trader voice CDRs when using absolute segmentation timer.

Added zone information to turret- and trader-level alerts.

The recorder can now create optional CDR/CTI records for the push-to-talk section of calls. These records can also be exported.

Media file download for Zoom Meeting/Webinar/Phone archiving now automatically retried 3 times on failure, and for Meetings/Webinars a checksum is validated on the downloaded file.

Archiving (importing) Zoom webinars support added allowing automatic archiving of Zoom Webinars for regulated users. CDR reconciliation is also supported for Zoom Webinars.

jQuery library is updated to 3.6.0

Apache Tomcat version is upgraded to 9.0.71, APR 1.2.36 (OpenSSL 1.1.1t)

New Role permission: once a user starts a controlled recording and then stops it, they should be blocked from starting recording again

Added the new "resource" parameter to the OIDC Identity Provider settings

Enable SSO Login without providing the Login ID in advance

The Web Application user session memory consumption was reduced

Reference Environment Supervisors are now able to view the metadata field values of other Environments.

The Legal Hold All API has been renamed to Delete Disabled API and it is now multi-tenant aware.

When a 2N controlled recording is started or stopped using the web application, the application now looks for the corresponding 2N recording pair, and also starts or stops that recording automatically.

The Conversations REST API now also returns the start and end times in GMT, in addition to the local time zone set for the user.

Added Select/Deselect All buttons to the Import Teams/Channels feature.

Added rewind and forward functionality to the player.

Improved UI translations for German, French, Japanese, Korean, and Russian languages.

Optimized the search performance for Group Supervisors that have many supervised users.

Added new permission "Start/Stop Controlled Recording" for the Controlled recording Start/Stop feature.

New Users Advanced IM Summary report showing the number of messages on a daily basis for every participant in a Microsoft Teams chat or channel.

New Conversations Advanced IM Summary report showing the total number of messages for a Microsoft Teams chat or channel

New Report: Advanced IM Export Policy Summary Report. The Advanced IM Export Policy Summary report shows the number of emails sent from VFC Capture to SMTP export targets and includes time stamp information about conversations. The report lists the number of successfully sent and failed emails, the start and end times between which chat messages were sent in the conversation in UTC and the user's time zone, and the target group filter for each export target.

Two new reports have been added: - Advanced IM Export Users - Advanced IM Export Count

Performance optimizations for Legal Hold checks for database maintenance job, and Legal Hold API to handle large volumes of data

Decreased the Legal Hold status update batch size from 250 to 50 to minimize the possibility of a database table lock.

Added a delay between Index Rebuild database operations to prevent the Transaction Log growing. The default configuration is that the operation sleeps for 15 minutes every 60 minutes.

Daily Jobs now avoid opening a database transaction when acquiring a database lock.

Internal domain patterns can now be configured on a per environment basis in multi-tenant mode. Available for Microsoft Teams voice/video/screen recording integrations.

The configuration of internal domain patterns on a per-environment basis in multi-tenant mode is now also available for the following services: - Unified Recorder Service including Cisco JTAPI integrations, - Cisco Compliance Service, - Genesys CTI Service, - Avaya DMCC Service, - Skype for Business (SfB) filter, - SfB/Lync IM Recorder Service.

Replay Only License Enforcement (93-530-3153, 93-530-3154, 93-530-3155)

Added four new Microsoft Teams bot-related alerts to the SCOM MP and SNMP MIB files.

Hitachi Content Platform Storage Target: improved error retrying mechanism to avoid the delay for the action as normally it was only retried in the next policy execution

Deduplication skips 2N records under Legal Hold.

Added support for S3 storage targets (S3 AWS, S3 IBM, S3 compatible) to the migration of recordings from Verint to VFC.

Added Microsoft Teams voice, video, and screen deduplication support for 2N recording.

CDR metadata is now generated for non-IM interactions during policy based export.

Policy-based export: if CDR metadata is not available for voice/video recording, it is now generated based on data available in the database.

Added media folder setting to the PowerShell Deployment Toolkit.

The VFC Installer now uses AES256 encryption for the private key when generating a new SSL certificate. Existing keys are checked, and converted to be using the AES256 encryption.

Added new Avaya Contact Recorder (ACR) Import Source which allows migrating historical recordings from an ACR system. The ACR Import Source migrates both ACR-only and ACR+15.2 systems using the native ACR API. Both encrypted and unencrypted ACR recordings can be migrated by the ACR Import Source.

Import Service now creates separate log files for each import policy, similar to the Storage Management Service. You can find the log files under Verba\log\import policies\. The same log settings apply to them (log file count, size, etc.).

Running the same import source in a multi-server environment could lead to false-positive alerts.

Genesys Cloud CX is now available as an import source.

Cisco Webex Meetings is now available as an import source.

SPAN-based SIP recording: In rare cases, a TCP reassembly bug could have caused the skipping of important SIP messages, leading to the related call not being recorded.

The method to determine Skype for Business call directions when no Internal Numbers/Domains Patterns were set did not work in VFC version 9.8.5.

Improved message error handling between the Unified Call Recorder and the Cisco JTAPI services to handle recording control command cache writing failures

Cisco Network Based recordings were stuck as ongoing when the recording service only received a very small amount of RTP packages.

Passive recorder could crash if in a RED-encoded audio stream it also received DTMF/telephone-event packets. This scenario could only happen if the call does not use Silk codec and switches to the redundant payload mode due to network reasons.

Multiple empty database entries were created for the same group chat when multiple Recording Serves were configured. The issue only affects v9.7.2 or later versions.

Recorded IMs did not store not recorded users in the participant list for scheduled meetings and Meet Nows.

Microsoft Teams Custom Announcement was not played for federated calls when the UPN of the federated party could not be retrieved (in most cases)

False positive one-way media error might have been detected for Microsoft Teams conference calls when unmixed (stereo) recording was configured

The Apply User's Recording Announcement service configuration turned off the recording notification entirely instead of turning off the user level announcement configuration only.

A false Call Processing alert was raised by the Unified Call Recorder Service intermittently when the Microsoft Teams Bot Service updated call information after the calls ended

Microsoft Teams service displayed a false positive "recording has stopped" banner if a peer to peer calling scenario was escalated to a meeting

The audio-only fallback implementation utilized to short waiting time and the waiting for the first participant list was not cancelled for terminated call legs. Therefore the bot made incorrect records.

When recording meetings with a large number of participants, Microsoft Teams calls were sometimes recorded in multiple segments instead of one.

In case of a screen share only recording scenario that was escalated to an audio call with audio recording only mode, the recorded call started from the beginning of the audio call escalation and skipped the silent screen sharing interval.

Bug fixes to the Microsoft Teams Application: - If the primary recording server was not available in a 2N recording scenario, the Microsoft Teams Application did not show secondary ongoing calls. - Supervisor users could only see their own calls, not the calls of others.

Internal calls that were originally not recorded got recorded after a hold/resume event.

Microsoft Teams Bot service failed when the organizer of a meeting was a third party application instead of an actual user.

In rare cases, audio recording can start even if the voice modality is not enabled.

Enhancement to mixed controlled recording scenarios: if a recording is triggered by an external participant, the recording becomes always-on recording until the last external participant leaves the call. If no external participant remains on the call, the recording continues, but the recording becomes controllable again.

In the case of call chaining scenarios, the PSTN direction was incorrectly determined as federated direction.

In a consultative transfer scenario with call chaining, the call leg of the transferring party inherited the recording configuration of the user who answered the transferred call.

In rare cases, when a recorded user who had call forwarding set up to a PSTN user received a call from a not recorded user, and the PSTN user answered the call, the call was classified as internal, and no recording announcement was made.

The Microsoft Teams bot service failed to answer a call if there was a not recorded third party on the call and the recording policy was configured as strict.

Multiple issues resolved for Microsoft Teams chat and channel archiving: - Attachments could not be archived (download from Sharepoint) if the user shared a file that was initially shared by another user - When the first message was edited/deleted or reaction added, the system created the CDR record with the wrong start date and time when the Export API based integration was used - If due to any reason, the UPN of a participant could not be retrieved and the participant entry was created without a UPN, it was never updated later, even the UPN could have been retrieved successfully - Media-Only records were not always appropriately linked to CDR-Only records in the case of chats for Export API based Microsoft Teams integration - From/From Info fields did not always reflect the recorded participant information in the CDR for Export API based integration - Channel, chat and participant information could have become inconsistent if multiple Recording Servers were deployed and multiple servers processed the same chat/channel. - Possible Unified IM Recorder service crash due to a race condition in the internal participant cache - DLP / Webhook API system event based channel name update was not updated on all records if the name was updated on a channel where the participants were not yet determined - The Unified IM Recorder service could have crashed using the DLP / Webhook based integration for Microsoft Teams when at service shutdown the webhook subscriptions were not properly terminated - Consumer Skype user was not properly recognized as a participant and generated multiple participation entries - The local cache polling mechanism (which moves the batch files to the global queues) was too eager and in some cases caused high CPU usage

Join/leave system events for private and shared channels could trigger join/leave events for all channels in the given team in the system internally. This could cause unnecessary system event entries, membership records or CDRs, and in certain cases data loss. The issue only affects DLP / Webhook based integrations, Export API based systems are not impacted.

Unified IM recorder service is now able to properly detect stuck message processing state and recover from it.

Changed the periodic channel and group details update from synchronous method to asynchronous to avoid a message processing deadlock caused by the Microsoft Graph API throttling.

During the export of Microsoft Teams IM messages with attachments, the attachment references were only added to the export of the original message. The exports of any edits or reactions to the original message were missing the attachment references.

Export API: In rare cases, edited chat messages were not exported. This issue only occurred if a chat message was edited and no newer message was sent during the export segment time period.

Made various improvements to the Microsoft Teams IM recording service to reduce the number of unnecessary API calls, such as revising processes for querying and caching chat details, and downloading attachments.

In rare cases, messages had incorrect timestamps during a recorder failover.

When a device was part of a call and then the call got transferred, then an unnecessary alert was sent because the system tried to record the device which did not even participate in the call

In rare cases, the Avaya recorder service inserted the wrong value in the From field after a transfer.

Unified Recorder: trader voice, CDR segmentation: PTT markers are inserted only for the last record but not at records terminated due to media segmentation. IPC Unigy and IPTrade integrations affected

IPTrade: The max retrieved call count for turret side playback is now configurable, and the default value is increased.

Ad hoc recording: Control sessions not dealing with media were included in the Overload state calculation based on the Max call count threshold. These sessions do not consume any resources, so after this fix, they no longer count towards the number of calls in the Recording calls statistics.

Call segmentation of Cisco Network Based recordings with JTAPI integration caused some extra CDR metadata that was received from the JTAPI to not show on the Conversation record.

Recordings of non-mixed Microsoft Teams audio streams contained undesired audio artifacts if the bot was configured to stream media with G.711 codec to the recorder.

When separate, not mixed audio stream recording was configured in interconnect call scenarios on Metaswitch Perimeta, the stream-participant association could fail for call participants with blocked caller ID, causing one-way media recording.

Race condition in Unified Recorder media recorder selection might have caused Recording Director service crash in very rare cases

IPTrade: if metadata becomes available only midcall (in certain flow device changes from unknown to defined one) the metadata was not updated at call termination

Midcall device changes were not handled (just ignored) if the same mix is referred in CTI. This behaviour was not transparent with other trader voice integrations. Now the recorder terminate the CDR and start a new one with the current device info as in case of other integrations

IPC Unigy Record on Demand: record CDR end cause was set to "normal" instead of "manual termination" when user disabled recording midcall

Media mix down alerts were incorrectly generated as media mix up alerts.

The Cloud9 CallData import source ignored the begin date and end date headers returned by the Recordings API endpoint.

The Cloud9 CallData API import source did not correctly filter metadata received from the API, which sometimes resulted in the call metadata having a shorter call length than the stored media. This caused the playback and export services to process the media with the shorter call length, creating a cut version of the media.

Maximum SIP session expires configuration was not taken into account by Cisco Announcement and Dial-in recording service causing calls to timeout after 2hrs if no session timer is negotiated.

Added a five second delay to wait before retrying an operation after an error to stop incorrect configuration generating excessive audit log messages that result in large audit log files.

In rare cases the native ID of a call was incorrectly shown in the To field of a conversation instead of the User ID of the called participant.

After a call hold scenario between two agents, the Genesys CTI service sent a start recording command for the same DN twice, instead of sending a command for each DN.

The Genesys CTI service could not create a CDR XML file for one of the call legs in scenarios with two agents.

A rare race condition in the media file download threads could result in 0 length or corrupt files with file checksum failures for files belonging to the same meeting/webinar.

Recurring export started from the search page is not allowed, and a warning is displayed about not supporting relative time intervals.

The "Update user information on existing conversations" feature did not work when the user had just one conversation

Azure AD Synchronization Profile Group selection didn't work in new Chrome versions

When the Reference Environment Default Conversation List Layout Configuration contained Metadata Fields, then a newly created Environment's default layout was corrupt and the Search would not work for the users created in the new Environment

When a user has got just Read Only administration permission, then he/she could list other environments' records

In Multi-Tenant mode, when an Environment's Active Directory Profiles deactivated a few Users or Extensions, then the number of deactivated objects was added to the subsequent Environments' summary message too

Searches for groups with "[ ]" brackets or "*" wildcard characters in the name did not show correct results.

Various reports generated 0byte files when a user filter was applied.

The Symphony platform was not selectable on the Extension configuration page.

Microsoft Teams IM conversations could not be listed on the Conversations search page after an Archive policy had been executed on them.

The Microsoft Teams IM PDF Export did not work when there was only a short activity in the selected conversation, and access scope of the logged in user was not set to All.

On the Conversation search page, highlights and labels from the previous search remained visible when the user jumped between pages, until the new highlights and labels were loaded.

On the Data Governance Dashboard, the default date interval of the Transcription Summary and Encryption Summary dashboard widgets was year.

Invalidated user remained visible in the "User most recent conversations" report.

Microsoft Teams IM records were returned in Conversation search results, even if the Instant Messaging option under Text Search was not selected.

The Peak Concurrent field of the Server Capacity report was not populated.

The following reports displayed data from other environments in multi-tenant mode: - Advanced IM Export Policy Summary, - Export Summary, - Extension Configuration, - Profile Configuration, - Server Configuration, - User Retention Details, - Users Without Any Recording.

The Exports dashboard widget displayed export tasks from other environments in multi-tenant mode.

The device name- and IP address-based Cisco phone service URLs did not work and threw errors.

Announcement cards were not displayed in Microsoft Teams IM messages. Affected release version 9.8.4.

The Delete button was disabled on the Data Management Policy Configuration page for Advanced Export policies, and those configured policies could not be deleted from the system.

The User and Group filters under the Conversation detail record fields Advanced Search option were ineffective for conversations that were shared through labels.

Users now have access to their own Ongoing Calls with the List Ongoing Conversations permission, even if their Conversation Access Scope is set to Access Shared Only.

Playback of shared calls did not work when the Conversation Access Scope of the user was Access Shared Only and did not have the Access View Shared Items Menu permission.

Fixed SQL injection vulnerability in Reporting Module.

Users Roles, Permissions, and Groups report generated an invalid file when filtering on a user.

Users Instant Messaging Details (Basic) report generated an invalid file when filtering on a group.

Simultaneous Calls Trend report generates an invalid file when filtering through multiple time zones.

Transcription data couldn't be inserted into the database. Every transcription provider was affected except Intelligent Voice. Affected version: 9.8.0.8036.

When the Legal Hold API was used to remove a Label from certain calls, then all other labels were removed from the calls too

Automatic Labeling Rule processing could have blocked other processes and so Upload could have been delayed when large volumes of trader voice records were inserted and matching the labeling rule. In addition, the Automatic Labeling Rule processing will be delayed by 60 minutes (was 10 minutes previously) and only process records updated 60 minutes ago.

Upgrade script for v9.7.7 caused duplicate Verba Maintenance jobs (named "Verba Maintenance Job (DB_NAME)" and "Verba Daily Maintenance Job (DB_NAME)") which could have caused maintenance job errors. As a manual workaround, the duplicated jobs can be removed. The "Verba Frequent Maintenance Job" must not be removed in any case.

Imported MPEG4 audio files without duration information in the media header could not be processed correctly.

SNMP trap target port configuration did not work. SNMP trap source port is no longer configurable.

Transcoding did not work for encrypted conversations, because the signing process failed after transcoding the media files with the following error message "Transcoding error at call #3a4a046f-002d-4278-a457-f941a40a548d: Read/write mismatch to encrypted file..."

Re-encryption did not work for files stored on non-SMB storage targets. The process did not return an error and the files were kept encrypted with the old key.

Microsoft Teams chat/channel archiving export (SMTP) issues resolved: - Microsoft Teams chat/channel SMTP export did not handle special characters for external users without a UPN and during SMTP export the email address of the external users were not complete - Possible race condition in the file and attachment download/export cache if the same file was referred by multiple threads/multiple conversations processed on different threads paralyze. This affects Microsoft Teams DLP / Webhook API based integrations and trader voice integrations. - If one message contained multiple attachments, and at least one of them was not available, the attachment links in the email message were not accurate - Failed email export if the FROM header contains and uppercase character in the email address - Failed email export if there is a single message in the export and the message time and the user participation start time were rounded to same msec by the SQL Server - [StartTimeInUserTimezone], [EndTimeInUserTimezone] fields in the SMTP export target configuration did not display the timezone information correctly - Attachment file extension was missing for images sent from an Android Teams client - Custom image attachments (e.g. OS specific sticker image) sent from an IOS Teams client could not be exported

When retention was set on an object using AWS S3 governance mode, the system could not increase retention. This issue happened when the upload policy did not set any retention, and the bucket had default retention using governance mode.

If standard instant message (SfB, Symphony, Cisco, etc.) IM transcript file was generated and uploaded with encryption, then export failed

Verint migrated call playback in rare cases could not work when the storage service couldn't identify the suitable archived version of the call.

Media mixing related Voice Quality scoring in rare cases was not properly calculate, causing false positive media mixing error flagging and low score

Policy based upload: race condition between record to DB insertion vs upload policy evaluation rarely might lead to match a lower priority matching policy instead of the expected

Legal hold: If multiple Media Repositories processed the same legal hold action, one of them could fail and trigger false-positive alerts.

Media stitching: Corrupted CDRs that are stuck in an ongoing state could cause abnormal call lengths with added silence in stitched media, leading to excessively large media files at playback or export.

SMPT export failed if the extension contained the apostrophe character.

Background task entry could not be created or updated and database down alert was triggered when advanced im export policy processed time ranges where given conversation/chatroom did not contain any message. Once policy finished, the background task entry with correct finished/failed counters was created.

SMTP export email address validation incorrectly removed the following special characters: (!#$%&'*+-/=?^_`

In Advanced IM Export, if only one of multiple attachments in the same message failed to be retrieved, the related export message did not properly replace JSON or structured attachment information with a user-friendly description of the attachment.

The S3 storage implementation policy could add log messages to the log file of unrelated storage policies, leading to incorrect and misleading log entries.

Improved the MP4 transcoding performance for video calls: - Transcoding is now 10% faster. - Video encoding profile bitrate is kept more precisely. - Better lip-sync. - Better stability.

In rare cases, the Move Media policy failed when moving Microsoft Teams IM attachment files.

The Storage Management Service kept crashing if an earlier upload process did not complete.

Media files downloaded from an Azure Storage target were not deleted from the local cache if the file integrity check failed. This caused excessive disk space usage.

During Policy-based export for non-Instant Messaging integrations, the VFC-generated CDR XML data was incorrect if the original XML was not available on storage. This issue affected every related integration, except for Verint-imported calls.

S3-compatible storage authentication failed when the REST endpoint of the storage target contained the S3 bucket name.

The upload to Azure Blob storage was forced to set the blob-level retention, even if the container did not support it, leading to upload failure. Only affected VFC version 9.8.5.

The Advanced Export always put .im files in the root folder instead of the YYYY/MM/DD folder.

Digitally signed calls could not be exported if the certificate used for signing expired and either the decryption was set or the calls were decrypted by a Sign & Encrypt policy.

Imported media files in MP4 format were not returned by the Media Streamer.

In rare cases, the upgrade to version 9.7 or above may be failed due to an SQL script execution error. The issue only affects the deployments which are using a custom metadata template that has a configured field with an empty Field Identifier value.

Certification generation failed during the installation of an additional server to a deployment that already had an installed server.

Legacy OpenSSL certificate generation step has been removed from the Additional Server Role installer.

Verint WFO Migration record counting can be extremely slow if a previously counted subset contained only a few records

Import service encountered in an error during the shutdown process

When an import policy finished, the related background statistics values were set to zero. Affected the release version 9.8.2.8149.

In rare cases, importing from Azure Storage could fail due to improper HTTP HEAD request handling. Affected release versions 9.7.7.7871 and later.

Improved the speed of VFC import source processing from Azure File Share. An additional file integrity check has been introduced.

The Import Source deduplication filter allowed the same conversation to be imported multiple times when the recorded_cid and agent_id fields were missing from the CDR XML file.

Migration from Verint WFO v11 did not work if a WFO User ID contained non-numeric characters.

The Offline Player did not open VMF files if the file extension was in upper case format.