Product:
Category:

February

O Flussonic Media Server 24.02 oferece novos recursos, incluindo compartilhamento de tela privada para o WebRTC Player mostrado em nossa demonstração em demo.flussonic.com. Aproveite o streaming contínuo para redes sociais diretamente das câmeras IP e a eficiência aprimorada do DVR com sondagem periódica da fonte M4F, entre outros aprimoramentos. Leia os detalhes sobre esses novos recursos e correções adicionais abaixo!

New

  • #9205 (Flussonic) Added the ability to add a silent audio track to video-only streams like from IP cameras with output_audio add_aac. This allows such streams playback in an environment requiring audio, such as social media or some TVs.
  • #9572 (Flussonic) Added support for periodical M4F sources polling for DVR info even for inactive On-demand streams. This allows for traffic savings due to eliminating the need for live on-demand stream activation when requesting its archive and reduces zap time.
  • #18470 (Flussonic) Added full HTTP/2 support for all endpoints including the Streaming API and API v3. This improves the Admin UI performance and allows reducing CPU load by reusing the connection via HLS and DASH.
  • #22967 (Flussonic) Added the support for pushing SRT with constant bitrate which is useful when receiver software or hardware requires CBR.
  • #33388 (Coder) Added the integration with Streamlabs USB-IO boards for switching stream inputs by pressing hardware buttons.
  • #33979 (Flussonic) Added the ability to use Opus codec in LL-HLS output.
  • #34065 (Flussonic) Added the ability to get events from Beward intercoms via ONVIF.
  • #35513 (WebRTC Player) Added the support for resuming playback when changing the stream’s input. You can change the phone screen orientation (landscape or portrait) or switch from publishing input to file backup and backwards while enjoying smooth continuous playback.
  • #35812 (WebRTC Player) Added the ability to send MediaStreamTracks directly to Flussonic WebRTC Publisher using the inputStreams parameter of WebRTC Player. With this feature, you can let the caller select a publishing source (web camera, microphone, screencast, etc.) before publishing start.
  • #35796 Updated the general pages for Flussonic Media Server and Flussonic Watcher with a comprehensive list of their features.
  • #36189 Pages from the “Developers” section divided into other sections in accordance with their content.
  • #36497 Improved the description of Bearer Auth and Basic Auth in the API Reference to help you understand which data is needed for authorization and where to find them.
  • #36820 Updated the PDF version of our documentation to improve formatting, readability, and navigation.
  • #37012 The articles inside the solution sections were grouped to follow the solution’s pipeline.

Improved

  • #28858 (Flussonic) Our Erlang implementation of SRT is now enabled by default for all cases where SRT is in use.
  • #34156 (Flussonic) Our Erlang implementation of SRT is now enabled by default for publishing and playing. You do not have to change any settings to use it as the transition is completely seamless. With that new implementation, we will be able to solve the SRT issues faster and more efficiently. Feel free to contact us if you encounter any troubles.
  • #34480 (Flussonic API) Implemented the use of resource_id parameter instead of meta drm_id in the config file to avoid the ambiguity of the meta parameter.
  • #34996 (Flussonic) Improved the CMAF packetizer to make the first start of LL-HLS streams playback faster by avoiding unnecessary buffering.
  • #35647 (Flussonic) Improved the info about GPUs in Prometheus: the device_id is now returned among other OpenMetrics parameters to distinguish between the boards and facilitate their monitoring.
  • #35717 (UI) Implemented the use of resource_id parameter in the UI instead of the obsolete and ambiguous meta parameter.
  • #35809 (Flussonic) Improved the stability of our implementation of SRT.
  • #35815 (WebRTC Player) Updated the description of onEvent() to list it as an option instead of method in the readme.
  • #35879 (DVR Player) Improved the timeline zoom in DVR player to allow faster positioning and navigation through the archive.
  • #35982 (Flussonic) Improved the logging of ad insertion process to facilitate troubleshooting.
  • #36043 (Flussonic) Improved the updater logging to hide messages about absence of errors.
  • #36192 (Flussonic) Flussonic now returns a proper error instead of silently ignoring the JPEG logo uploaded for transcoder. You should use only PNG for overlaying logo with transcoder, while the other ways of adding logo allow the use of other formats.
  • #36396 raspberry pico as a GPIO USB board for playout server to work with external buttons
  • #35633 Added the explanation of the license use when reinstalling OS or changing the server.
  • #36030 Updated the article about MPTS ingest to explain the cases where MPTS is used and make the instruction more clear.
  • #36183 The menu on the left was regrouped to remove the confusing “Products” section.
  • #36187 Added the high-level “IP cameras” menu group for video surveillance solutions based on Media Server.
  • #36188 All pages related to Central were removed from the documentation as providing no value to our customers.
  • #36197 Administrator’s instructions moved to the “Manuals & Guides” section.
  • #36200 Removed the parameter descriptions (mostly invalid) duplicating those in the API Reference. Remember that API Reference is always the most relevant and reliable source of information about any settings.
  • #36506 Added the “Manuals & Guides” section in the main documentation menu on the right. This section includes step-by-step configuration instructions.
  • #36868 Fixed the recommendation on configuring NTP for sending multicast to make it clear that Flussonic server should synchronize time with NTP server.
  • #37071 changed location of md files from two similar directories to one directory with multilang files
  • #34387 Added the demo for Flussonic’s implementation of Low-Latency HLS (LL-HLS). Check it out at demo.flussonic.com.
  • #34750 Added the demo for screen sharing with WebRTC Player. Check it out at demo.flussonic.com.

Fixed

  • #9322 (Flussonic) Fixed the issue with archive clean-up not performed when global DVR config is set for cloud storages like Amazon S3 or Swift.
  • #27011 (Flussonic) Fixed the artifacts when playing H.265 (HEVC) encoded RTSP caused by incorrect Fragmentation Units.
  • #32143 (Flussonic) Fixed segfaults appeared when restarting the stream received from a timeouted Dektec SDI source.
  • #32819 (Flussonic) Fixed the issue with WebVTT subtitles disappearing in output stream after a short failure in the source with DVB subtitles.
  • #33659 (Flussonic) Fixed the issue with spikes in CPU consumption that may have occurred when using a Guardant USB key for licensing. The use of Guardant does not affect the service any more.
  • #34053 (Flussonic) Fixed the issue with #EXT-X-PART HLS segment duration exceeding PART-TARGET which was reported as must-fix error by Media Stream Validator and could prevent the stream from being played.
  • #34586 (Coder) Fixed the issue with the changes to static IP address not saved when configuring the streaming0 interface in the UI.
  • #34663 (DVR Player) Fixed the issue with DVR Player referring to Watcher API v2 instead of proper episodes URL to get data on episodes.
  • #34907 ASI output now has 100% PCR accuracy and zero PCR jitter
  • #35352 (Flussonic) Fixed the incorrect counting of clients for On-demand streams which resulted in watched stream stop.
  • #35744 (Flussonic) Fixed the issue with stream ingest not performed when the protocol name was specified in upper case (for example, HTTP).
  • #35762 (WebRTC Player) Removed the obsolete onWebsocketClose() method from the readme.
  • #35861 (Flussonic) Fixed the issue with negative value for EXT-X-MEDIA-SEQUENCE in HLS/EVENT DVR playlist.
  • #35890 (UI) Fixed the issue with “Input media info” not shown in the UI if any input except the one with the highest priority was active.
  • #35892 (UI) Fixed the issue with the player on the Overview tab not playing video when a custom token is used for session authorization.
  • #35899 (MSE-LD Player) Fixed the issue with MSE-LD playback freezing when switching back to video+audio track after an audioless track in embed.html player.
  • #35977 (Flussonic) The memory allocation mechanism has been tweaked to save resources and avoid out of memory errors.
  • #35979 (Flussonic) Fixed the issue with the copy:// source for EIT in MPTS copying only one program’s EPG.
  • #36024 (Flussonic) Fixed the issue with LL-HLS playback of on-demand streams interrupted even if there were clients watching them.
  • #36172 (Flussonic) Fixed the issue with Flussonic SRT authorization and connection errors returned to the Caller (player) as unknown errors without proper descriptive messages.
  • #36178 (Flussonic) Fixed the issue with the number of active play sessions incorrectly displayed in the UI.
  • #36179 (Flussonic) Fixed the issue with incorrect SSL flags in playback sessions.
  • #36276 (Flussonic) Fixed the issue with DVB-T2 ingest not performed because of unnecessary T2 MI decapsulation.
  • #36390 (Flussonic) Fixed the issue with trimming when pushing CBR UDP from an uneven source.
  • #36765 (Coder) Fixed the issue with incorrect saving of static IP address on manage0 interface.
  • #36816 (Flussonic) Fixed the HTTP 500 error returned by recording_status.json request if an invalid archive source is specified.

Bem-vindo ao Flussonic Watcher 24.02, onde estamos transformando sua experiência de monitoramento! Mergulhe em uma visão mais completa da câmera com os mais recentes tamanhos de mosaico introduzidos nesta versão. O recurso de mosaico foi projetado para aprimorar o monitoramento eficaz de suas instalações.

New

  • #15188 Added the support for 5x5, 6x6, 8x8 mosaic sizes in Watcher API v2.
  • #21775 You can now add up to 64 cameras to the mosaics in Watcher UI. Just choose the size in the mosaics editor: 5x5, 6x6, 7x7, or 8x8.

Improved

  • #33313 Improved the support for cameras that provide slow connection via ONVIF.
  • #35777 The network statistics in the streamer list in Watcher UI is now enabled by default.
  • #35821 Refactored the mechanism of instant analytics events to use the Long Polling technology. The analytics messages can be used for receiving push messages in Watcher mobile app or for integration with access control systems.
  • #35912 Added the ability to stop, start, and restart flussonic-watcher service with all its dependencies. This is required to stop Watcher for creating a consistent backup.
  • #36052 Removed the storage of Agent metrics in Redis and Prometheus monitoring for them as those functions were excessive and needless. Abandonment of Redis, when finished, will facilitate the Watcher use in a cluster or container orchestration system like Kubernetes.
  • #36233 Improved the indication of authentication error when connecting to the camera via ONVIF.
  • #36387 Simplified the Vision module installation by eliminating the need to add the NVIDIA repository.
  • #36627 Fixed HTTP 403 error when embedding cameras extra protected by authorization with token.
  • #36798 Watcher installation wizard was improved to support the /central/api/v3 API endpoints.

Fixed

  • #34032 Fixed the issue with updating XM firmware from Watcher UI. Please note that you should update both Watcher and Media Server for the fix to be available.
  • #35506 Fixed the process of binding persons and episodes with consideration of Organizations. If added to several Organizations, the person is updated only in the same Organization as the camera where it was detected. When a person is recognized on a camera in an Organization to which the person is not added, the event is not raised.
  • #35540 (UI) Fixed the issue with branding logo in .svg format not displayed in Watcher Client UI (PWA).
  • #35716 (UI) Fixed the issue with the position in camera list dropped when opening a camera and then returning back to the camera list in Watcher Client UI (PWA).
  • #36033 Removed the NVR config provisioning from Watcher as it is now configured on Central side.
  • #36182 (UI) Fixed the issue with crashing ONVIF settings tab in the camera profile in Watcher UI.
  • #36295 Fixed the issue with camera folders and Organizations reset to default (assigned at first provisioning) when reprovisioning the same cameras with Agents.
  • #36307 Fixed the issue with the disabled cameras displayed as online in the UI.
  • #36556 Fixed the issue with signing into Watcher Client UI (PWA).
  • #36591 (UI) Fixed the issue with person get request not executed when you select a person in the list at Watcher UI.
  • #36603 Fixed the issue with displaying that push notifications are enabled for a user who did not enable it if that user logged into Watcher Client UI after the user who enabled the notification.
  • #36619 Fixed HTTP 500 error when saving a person in Watcher.
  • #36770 Fixed the issue with Agents not provisioned to Central when imported to Watcher.

Conheça as notas de versão do Central 24.02, onde implementamos recursos aprimorados de controle e monitoramento para a análise do Central, além de vários aprimoramentos e correções de bugs.

New

  • #34906 Implemented the display of the number of online streams on analytics nodes.
  • #35516 Supported the layouting of streams with analytics enabled between analytics nodes only; failover and stream limits are also available.
  • #35827 The status and last_detection_at data are now collected for the streams with analytics to facilitate troubleshooting.

Improved

  • #35182 Improved the stream layouting to speed up stream allocation at the first launch or restart of Central by immediately launching the layouter.
  • #35811 Implemented liveness_probe and readiness_probe methods allowing you to check whether Central is started and ready to use. This is helpful when deploying your cluster in Kubernetes or other container orchestration system.
  • #36436 Added the ability to get stream.vision.stats from Central.
  • #36551 Improved the events filtering by camera name.
  • #36624 Added the ability to set the stream’s layout and the batch stream save operation for setting the layout for a bunch or all streams.

Fixed

  • #35993 Fixed the issue with streaming endpoint hostname not assigned to Agents.
  • #36494 Fixed the issue with NVR Agents resetting unexpectedly some time after activation.
  • #36547 Fixed the issue with stats for identification nodes not saved leading to the analytics events not shown.
  • #36616 Fixed the issue with the stream stats returned for disabled streams making them appear as online/alive. The stats are now cleared when disabling a stream.
  • #36760 Fixed the issue with streams allocated to analytics nodes while not being ingested on any streamer.
  • #36761 Fixed the issue with person previews not saved for persons created by the analytics module making it impossible for the user to identify the person.
  • #36832 Fixed the issue with Agent disconnections caused by the streampoint key resetting after streamer update.
  • #36843 Fixed HTTP 500 error when requesting the Agent update.
  • #36847 Fixed the issue with streams not allocated to streamers when created from Watcher.
  • #36849 Fixed the issue with episode previews not updated when the episode was updated by an analytics node which resulted in low quality of the preview in Central.
  • #36861 Fixed the issue with Central not returning stats for analytics nodes.
  • #37022 The Central’s environment variables now have the CENTRAL_ prefix.

January

O Flussonic Media Server apresenta sua última versão, 24.01, um modelo de inovação em streaming. O primeiro changelog do ano traz entusiasmo com demonstrações dedicadas que o convidam a explorar os extraordinários recursos do Flussonic. Mergulhe no futuro do streaming com WebRTC para videoconferência, publicação SRT com uma porta dinâmica e o intuitivo DVR player, todos disponíveis para teste em demo.flussonic.com.

Embarque em uma jornada de descoberta e inovação. Experimente todos esses recursos gratuitamente. Explore, experimente e testemunhe os recursos de ponta da versão mais recente do Flussonic.

New

  • #15418 (DVR Player) Supported LL-HLS playback in embed.html. You can use ?realtime=true&proto=ll-hls query string to request it from the server.
  • #24926 (Flussonic) Implemented a framework for appliance firmware creation.
  • #26826 (Flussonic) Supported FairPlay encryption of HLS HEVC fMP4 streams for all DRM providers.
  • #32762 (Flussonic) Flussonic can now accept publications from Cisco CMS via RTMP.
  • #34074 (DVR Player) Finalized the implementation of the archive fragment downloading from DVR Player. You can check it out in the DVR Demo.
  • #35454 (Flussonic) Supported DVB subtitles conversion to WebVTT for SRT published streams.
  • #35517 (Flussonic) Added the ability to use several words divided by space in push or input descriptions.
  • #35576 (Flussonic) Implemented the ability of Flussonic Media Server to proxy requests to Central installed on the same server. This feature can be used in single-server Watcher installations.
  • #35623 (Flussonic) Our own SRT implementation (just like the previous one based on libsrt) supports key refresh which is important for long-time broadcasts.
  • #35706 (Flussonic) Added the config_external_success event indicating that the configuration backend is requested successfully.
  • #33203 The article about creating your own IPTV channel was updated with instructions on ad insertion.
  • #34899 Fixed the example of Prometheus configuration to include a more clear targets example.
  • #35155 Demo WebRTC call is available at demo.flussonic.com. It demonstrates the Flussonic WebRTC capabilities in a video chat room.
  • #35360 Added the demo of SRT publishing with dynamic port at demo.flussonic.com. The port changes every time you update the page. You can push your stream via SRT from OBS Studio, Larix Broadcaster, ffmpeg, Flussonic Media Server, etc. The stream instantly shows in the player.
  • #35733 Added the demo of DVR Player at demo.flussonic.com. You can watch live video and archive of a CCTV camera and download the archive fragments.

Improved

  • #33851 (WebRTC Player) Updated the readme for Flussonic WebRTC Player npm package to add missing descriptions of methods like screenShare() and options like start_track.
  • #34460 (Flussonic) Improved memory allocation to enhance general system performance.
  • #35062 (Flussonic) Improved the H.265 (HEVC) transcoding to optimize CPU utilization.
  • #35087 (UI) The cluster key is no longer required when creating a streamer in Central Admin UI because you don’t need the cluster key for video analytics streamer roles.
  • #35208 (Flussonic) Improved our Erlang SRT Listener to support single port for play and publish in the same way as in the libsrt implementation.
  • #35698 (Flussonic API) Improved the requests routing with config_external enabled to bypass internal HTTP requests.
  • #35727 (Flussonic) Fixed the issue with RTMP push restarting every time the source is disconnected by timeout with backup file configured.
  • #35778 (Flussonic) Improved the archive cleanup process to exclude unnecessary operations.
  • #35789 (UI) Updated the name of the inference streamer role in Central Admin UI to make in clear that it should be selected when vision-inference package is installed on the corresponding streamer.
  • #35865 (Flussonic API) Moved srt_port_resolve to stream stats.
  • #36006 (Flussonic) Fixed the issue with unnecessary information in logs about disabled playback protocols.
  • #29042 Added transcoder instructions to the IPTV/OTT quick start page.
  • #31371 Added the article about the use of firewall and why you don’t need it on servers with Flussonic.
  • #33753 The Filtering collections principle of Flussonic API design was updated to clarify that _is and _is_not filters are only applied for comparison with NULL.
  • #34116 Updated the MSE Player page to remove obsolete information about use without npm and synchronous playback of several DVRs at the same time.
  • #34627 Described the option for copying EIT with EPG from a specified source to MPTS.
  • #34655 Fixed the link to API on the SRT page, fixed the description of the latency parameter used for SRT input and SRT push.
  • #35289 Added the sentence about various disk size allowed in Flussonic RAID for DVR.
  • #35465 LL-HLS Playback instructions moved to the separate page and updated according the latest changes in the API and UI.

Fixed

  • #31284 (Flussonic) Fixed the issue with ad_splice insertion when passing SCTE-35 markers through Flussonic leading to creation of too short segments that caused bad_segment message in logs and too long buffering.
  • #31993 (Flussonic) Fixed the issue with SPTS streams duplication when standby is enabled.
  • #32827 (Flussonic) Fixed the issue with about one second part of the file cut when transcoding it with file_processor.
  • #33995 (Flussonic) Improved RTSP playback stability for VLC and some other players.
  • #34023 (Flussonic) Fixed the issue with server playlist crashing when reading some files with specific subtitle timings.
  • #34200 (Flussonic) Fixed the freeze before injected ads in segments with subtitles.
  • #34345 (DVR Player) Fixed the issue with DVR Player showing black screen instead of the archive in the Dashboard mode in Watcher UI.
  • #34672 (Flussonic API) Fixed the HTTP 500 error returned by openapi_handler when required fields were missing in select query string request.
  • #34715 (Flussonic) Fixed the issue with transcoder failing and not restarting properly after a specific minor source failure.
  • #34748 (Flussonic) Fixed the issue with the actual ASI output bitrate being eight times higher than specified in the config.
  • #34820 (DVR Player) Fixed the issue with fast playback not starting without changing playback position on the timeline in Safari at iOS 17.1.1.
  • #34910 (Flussonic) Fixed the issue with Media Info changes causing stream restart.
  • #35084 (DVR Player) Fixed the issue with playback freezes in Chrome when selecting points on the timeline one after another.
  • #35109 (Flussonic API) Fixed the HTTP 500 error returned in response to PUT /transponders/{name} when PID bitrate is higher than multiplexer bitrate. HTTP 400 with pid_bitrate_higher_than_transponder_bitrate message is now returned instead.
  • #35231 (Flussonic) Fixed the issue with On-demand stream not starting when requesting its playback via RTSP.
  • #35305 (Flussonic) Fixed the issue with too often update checks causing excessive CPU load on the server.
  • #35345 (Flussonic) Fixed the issue with Media Server requesting the list of episodes from Central in a wrong way and being unable to delete the expired episodes recording as a result.
  • #35361 (Flussonic) Fixed the issue with update not performed from the UI.
  • #35383 (WebRTC Player) Fixed the issue with native controls in WebRTC Player not disabled by default and interfering with custom controls.
  • #35444 (Flussonic) Fixed the issue with exporting more than hour-long archive to an MP4 file on the server disk via the API interrupted by a timeout.
  • #35547 (Flussonic) Fixed the inconclusive error messages in logs indicating that IPTV plugin is not enabled or configured.
  • #35565 (DVR Player) Fixed the inability to scale and hover over the frame when the stream is paused.
  • #35691 (Flussonic) Fixed HTTP 500 error appeared when playing DVR from Amazon S3 over segment-based protocols when copying is configured and the segments that are not yet in memory are requested.
  • #35705 (Flussonic) Fixed the issue with archive cleaner reporting in logs about inability to delete already deleted archive blobs.
  • #35863 (DVR Player) Fixed the issue with numbers cropped in the archive fragment selection box.
  • #35876 (DVR Player) Fixed the issue with the “play” button displayed instead of “pause” when autoplay is enabled.
  • #35959 (Flussonic) Fixed the issue with SHOUTcast crashes leading to session interruptions.
  • #35566 Fixed the issue with main page my.flussonic.com opening instead of specific ticket when tapping a push message about a new answer in the ticket.

Na versão mais recente, o Watcher versão 24.01 surge como um modelo de inovação, introduzindo uma série de recursos que redefinem as capacidades de vigilância. Desde o escalonamento refinado da linha do tempo até as configurações otimizadas de análise de vídeo e a adição de um reprodutor DVR para o gerenciamento contínuo de arquivos, o Watcher tem o compromisso de aprimorar a experiência do usuário. A implementação de solicitações de proxy para a Central marca um salto significativo em direção a uma maior escalabilidade e otimização de recursos.

Experimente os novos recursos, aprimoramentos e correções de bugs e testemunhe a evolução do Watcher.

New

  • #35048 (UI) The timeline scale increased to show records for the last hour by default facilitating the archive navigation.
  • #33206 timelapse option is enabled for watcher ui users.

Improved

  • #31964 (UI) Deleted the “Vision” setting from streamer profile in Watcher UI. The relevant way of configuring the video analytics will be available in our documentation as soon as possible.
  • #35694 Supported the routing of API requests with /central prefix to Central.

Fixed

  • #30906 (UI) Fixed the issue with filtering on the Events page in Watcher UI resulted in emptying the Events page.
  • #33670 Fixed the issue with unrecognized persons not appearing in the event list and in the Lists of Persons.
  • #34121 Fixed the issue with two or three push notifications on one event in Watcher Client UI (PWA).
  • #34687 Fixed the issue with branding logo applying only after clearing the cache.
  • #34986 (UI) Fixed the display of events and player controls on the camera page of Watcher Client UI (PWA) when viewed in landscape orientation on mobile devices.
  • #35466 (UI) Fixed the issue with NVR cameras not played in the DVR Player when listed on a second or greater page in the camera list in Watcher UI.
  • #35601 Fixed the issue with Watcher not being able to connect with a camera via ONVIF when the camera responded with a minimum GetNTP response.
  • #35967 Fixed HTTP 400 error returned in response to GET /persons when persons in the old format were present in the database.

O Flussonic Central 24.01 marca uma grande atualização que introduz vários aprimoramentos importantes. A inclusão da tecnologia Long Polling garante uma entrega mais rápida de episódios, melhorando a eficiência geral do sistema. Otimizamos as conexões de câmera com a introdução do recurso Agent Assignment, proporcionando um processo mais simples e rápido.

Além disso, essa atualização prioriza a escalabilidade aprimorada com uma implementação mais otimizada de transmissores, o que é particularmente benéfico para ambientes orquestrados. Nosso compromisso com a otimização se reflete no tratamento aprimorado das impressões digitais faciais, o que contribui para o desempenho geral do sistema.

Em particular, a distribuição de episódios foi refinada, oferecendo recursos superiores para tarefas de análise de vídeo. Esses aprimoramentos, juntamente com vários outros, destacam nossa dedicação em melhorar a funcionalidade do Flussonic Central.

Para uma exploração detalhada dos novos recursos introduzidos no Flussonic Central 24.01, convidamos você a se aprofundar nas últimas notas de versão.

New

  • #31270 Enhanced the delivery of episodes to Watcher or other client application by using the Long Polling technique. This will help decrease the delay when using episodes for access control like opening door by a recognized face or opening gates by a recognized license plate.
  • #35196 Assigning Agent with a streampoint is made independable from the layouter to allow the camera connection even if the layouter is unavailable.
  • #35671 Supported the /central/api/v3/ API path along with the /streamer/api/v3/ to distinguish the services installed on the same server without configuring different ports.
  • #35918 Implemented the storing of updated_at in the Central’s database to avoid possible desynchronization and episode loss.
  • #36132 Added the ability to disable provisioning of configuration from Central to streamers. It is useful when you deploy your streamers in Kubernetes or other orchestration system. Set CLUSTER_NODE_CONFIG_PROVISION_ENABLED = false to use this feature.

Improved

  • #35259 Improved the distribution of episodes between the analytics nodes to allow Central push episodes batches for identification. This increases control over the results of episodes processing on the analytics nodes and provideы means for handling any errors that may be encountered during the episodes processing.
  • #35377 Improved the check for episode validity to forbid episodes with updated_at=0 and started_at=0 because such values do not make sense and may prevent validation by other services.
  • #35395 Improved the face fingerprint handling in Central to exclude merging data and utilize only the relevant data returned by the analytics nodes. This improvement increases the system performance.
  • #35402 Improved the validation of returned person list to exclude persons with invalid photos.
  • #35515 Renamed the name parameter to media to avoid confusion and highlight that the parameter has the same meaning as in GET /episodes.
  • #35597 Improved the episodes fetching from Vision Inference by implementing the polling technique to reduce the delay between the episode start and its appearance in Central.
  • #35617 Renamed vision role to inference to reflect that this role corresponds to vision-inference package.
  • #35696 Updated the config_external and central_event_collector URLs to use /central prefix instead of /streamer.
  • #36008 Default HTTP timeouts increased up to 60 seconds to improve Central operation in Watcher software package.
  • #36009 Removed the event sinks enabled/disabled setting because there is no need in disabling it.
  • #36040 Increased the system security by removing the password and API key generation when installing Central as part of Watcher software package. The users shall set the password themselves in order to use Central Admin UI.
  • #36044 Removed the SINGLE_INSTALLATION config parameter to facilitate scaling.
  • #36154 Central’s default port is now 9019.

Fixed

  • #35249 Fixed the issue with GET /episodes request with filter by updated_at returning not all the detections present in the database because the updated_at value was processed by Central in a wrong way.
  • #35265 Fixed the issue with streams moved to another streamer at failover not returning to the streamer specified in their settings after that streamer has restored after failure.
  • #35501 Fixed the issue with Central continuing to provision a stream to analytics nodes after vision is disabled on the stream.
  • #35571 Fixed the issue with updated_atsynchronization ensuring its uniqueness for every episode.
  • #35610 Fixed the issue with updated_at value not updated for the episodes received from the analytics nodes.
  • #35684 Fixed the issue with the streams having vision enabled not allocated to inference nodes.
  • #35922 Fixed the issue with scheme parameter not applied when executing GET /streams request.
  • #35923 Fixed the issue with analytics not working if cluster_key is not specified in Central which is wrong because cluster_key is not required for analytics nodes.
  • #36026 Fixed the issue with Agent statuses incorrectly probed by Central leading to incorrect expectations of the Agent allocation to streamers.