Skip to content

LUNA PLATFORM v.5.59.0#

LP changes

  • Note: In the next release, the default value of the neural network for extracting face descriptors will be updated from version 59 to version 62. Version 59 will be removed from the Remote SDK container.

  • A Storages utility has been developed that enables you to check and/or prepare the environment for LUNA PLATFORM v.5.46.1 and higher services before launching them.

    The preparation of the environment is understood as follows:

    • Preparing buckets in InfluxDB for monitoring work.
    • Preparing buckets for the Image Store service, enabling you to store user data (images, metadata, archives, etc.).
    • Preparing the InfluxDB for collection of aggregated statistics by the Admin service (see "Requests and estimations statistics gathering" section in the administrator manual).
    • Preparing databases, adding VLMatch functionality, creating and managing migration scripts for LUNA PLATFORM databases (PostgreSQL or Oracle).
    • Performing migration/loading settings into the Configurator database.
    • Loading dump files into the Configurator service database.

    The main advantage of Storages over manual preparation of the environment is that the utility knows revisions of the Configurator service settings and migration scenarios for LP service databases, which greatly simplifies the process of updating and downgrading LUNA PLATFORM. In addition, using the utility, you can evaluate the current state of the LUNA PLATFORM environment and determine what exactly needs to be updated.

    The utility is supplied in a Docker container and can be used as a tool for preparing the LUNA PLATFORM environment deployed in Docker containers, the Docker Compose script, the Kubernetes orchestration system, etc. The main task of the administrator is to indicate to the Storages utility the addresses of databases, buckets, etc.

    A new document "Storages utility manual" has been added to the package and to the online documentation site, describing the principles of using the utility and examples of use. Two new LUNA PLATFORM deployment manuals have also been added -"Install manual using Storages" and "Upgrade manual using Storages". The documents are similar to regular installation and upgrade manuals, except that all environment preparation is done using the Storages utility.

    Important: The Storages utility is in beta testing. See detailed information about the utility in the Storages utility manual.

  • SDK has been updated to version 5.18.0.

  • The ability to enable sending notifications about the status of subtasks ("subtask_callbacks" section) has been added to the "notification_policy" policy of each task.

    The subtask state contains the following information:

    • Subtask ID
    • Subtask status ("in_progress", "failed", "cancelled", "done")
    • Number of completed subtasks

    In responses to the "get task notification policy" and "replace task notification policy" requests also contain the information described above.

    Notifications about the status of subtasks are also available when using the "notification_policy" in the schedule.

  • A new resource /metrics has been added to all LUNA PLATFORM services, enabling you to collect and save metrics in the Prometheus format as time series data that can be used to track service behavior.

    By default, metrics collection is disabled. Metrics collection is enabled in the "enabled" parameter in the "LUNA_SERVICE_METRICS" section. If metrics collection is disabled, then a request to the /metrics resource will return an error with the code "12049" and the message "Forbidden, Resource is disabled".

    There are two types of metrics available:

    • request_count_total — Total number of requests.
    • errors_count_total — Total number of errors.

    Each of them has at least two labels for sorting:

    • status_code (or error_code for error metrics)
    • path — Path consisting of the request method and the endpoint route.

    If necessary, you can add custom label types by specifying the label_name=label_value pair in the "extra_labels" parameter in the "LUNA_SERVICE_METRICS" section.

    See "Export metrics in Prometheus format" for details.

  • The ability to obtain the X and Y coordinates of people when estimating the number of people has been added.

    The ability to obtain coordinates has been added to "sdk" request and the "detect_policy" of the "create handler" request. To do this, you need to enable the "people_count_coordinates" and "people_coordinates" parameters respectively.

    The generated event structure has also been updated.

  • Examples for the request "generate stream events (beta)" have been added to the OpenAPI specification.

  • The ability to run lambda GPU has been added.

    To use the GPU, you must enable the "deploy_parameters" > "enable_gpu" parameter in the lambda generation request.

    There are some requirements and limitations when using GPUs in Kubernetes. See "Create GPU-enabled lambda" administrator manual.

LP fixed errors

  • Fixed lack of support for the OPTIONS method for the resource /plugins.

  • Fixed the "Internal server error" that occurred when executing the Garbage collection task with the events filter enabled and the remove_image_origins parameter, when the event did not contain the source image (None value).

  • Fixed the value too long for type character varying(512) error that occurred when downgrading Configurator settings from revision d718e5954caf.

    Changes made in revision d718e5954caf led to an increase in the length of the token settings value string from 512 characters to infinity. When downgrading from revisions where the length value in the corresponding setting was already greater than 512 characters, the error value too long for type character varying(512) was returned.

    Now, when downgrading from revision d718e5954caf, a setting value exceeding 512 characters will be deleted. You must re-set the token value for the appropriate setting.

  • Fixed the error Key (name)=(<setting>) is not present in table "limitation" that occurred when downgrading Configurator settings from revision dd2641137b42.

    Changes made in revision dd2641137b42 led to the mandatory binding of each setting to a specific template. For this purpose, a new table limitation was added containing restrictions. The error occurred when performing a downgrade from revisions where the Configurator database already contained data about restrictions in the setting table and it was impossible to establish a connection between the limitation table being deleted and the existing setting table.

    Now, when performing a downgrade from revision d718e5954caf, all entries from the setting table will be deleted when the limitation table is deleted.