v.5.108.0#
Changes
-
The maximum length of the
reference
field (source stream address) in the create stream request has been increased from 512 to 2083 characters.Important: When downgrading to LUNA PLATFORM, ensure you have no critical streams containing links longer than 512 characters - these will be removed.
-
SDK has been updated to version 5.29.0.
In this version of LUNA PLATFORM DeepFake estimator updated to version 8.
-
The
face_descriptor_score
field has been added to the Remote SDK service monitoring to collect face descriptor quality data. See the "Monitoring" section of the developer guide for more information.The response to the "get system info" request for the Admin service has also been expanded: a new
face_descriptor_score
field has been added, providing face descriptor quality statistics. The data is automatically distributed into buckets with a step size of 0.05. Example:Here "score_0.9": 197 - means 197 descriptors with an estimate in the range 0.90–0.94."face_descriptor_score": { "score_0.85": 2, "score_0.9": 197, "score_1": 4 }
-
Monitoring data aggregation is now optional and can be disabled via the new
MONITORING_AGGREGATION
setting in the Admin service.Setting format:
boolean
. Default value:true
.Note: When the setting is disabled, statistics in the
stats
sections of the response to the request "get system info" will not be displayed. -
The logic of preparing the environment for monitoring using the Storages utility has been changed.
Previously, the creation of a bucket for storing monitoring data
luna_monitoring
(if no other name is specified) and the creation of a bucket for storing LUNA PLATFORM usage statisticsluna_monitoring_aggregated
(now optional, see above) were performed in two stages, and now this is a single stage of preparing the monitoring environment. This means that if theMONITORING_AGGREGATION
setting is enabled (enabled by default), during the preparation of the environment for monitoring, the errorBucket <bucket_name> already exists
will be returned if any of the two buckets exist, since the Storages utility perceives the absence of any of the buckets as the need to create both.In all installation docs and the Docker Compose script, the InfluxDB startup command is now executed with a bucket named
default
initialized instead ofluna_monitoring
(theDOCKER_INFLUXDB_INIT_BUCKET
environment variable for the InfluxDB container), and the Storages utility then creates the bucket(s) itself. This avoids an error when theluna_monitoring
bucket was already initialized during the InfluxDB startup step, and the Storages utility detects its presence along with the absence of theluna_monitoring_aggregated
bucket.If necessary, you can leave the usual InfluxDB container startup command unchanged and use the
--ignore-integrity
argument when preparing the Storages environment. This will create only theluna_monitoring_aggregated
bucket, ignoring the presence of the monitoring data storage bucket, but you should use this argument with caution, as it may cause errors. it can also ignore other existing entities when called with the positional argumentprepare all_entities
.Important: When upgrading from a previous version where the Admin service was not used and the
luna_monitoring_aggregated
bucket was not created, but theluna_monitoring
bucket was created, you will also get the errorBucket <bucket_name> already exists
. In this case, you need to disable theMONITORING_AGGREGATION
setting, or use the--ignore-integrity
argument (see above).
Fixed errors
-
Fixed incorrect output of the
luna_prepare prepare --help
command in Video Manager and Video Agent service containers. -
The support for the environment variable
EXTEND_CMD
has been restored for the Licenses service. With its help, you can transfer additional arguments to the service launch team, if separate variables are not provided for them.For example, you can clearly set the configurations tag when starting the service:
--env=EXTEND_CMD="--LUNA_MONITORING=TAG_1"
.See the "Service arguments" section in the installation manuals for details on environment variables and arguments.
-
Now, if the analyst is not available during the request of create stream, the status code 422 (Unprocessable content) is returned instead of 400 (Bad/incomplete input data).