v.0.34.0

Improvements:

  1. [LUNA-7660] Added an example of a standalone lambda with parsing multipart input with 2 images and json.

  2. [LUNA-7682] A new client has been added to lambdas, allowing it to interact with other lambdas through both lambdaId and lambdaName. For interaction via the lambda name, when it is created, 2 services are now created in k8s. You can, see examples here.

  3. [LUNA-8232] Added new setting LAMBDA_RESOURCE_LIMITS that specify available resources per namespace. Added new parameter check_resources to lambda creation, lambda update and lambda import requests. If this parameter is on, then before lambda deployment there will be check of requested resources from deploy_parameters field against limits from LAMBDA_RESOURCE_LIMITS setting. If reqested resources exceed limits error will be raised. Added new route lambda resources which shows available resources per namespace.

    See more at lambda resources

Bugs fixed:

  1. [LUNA-7682] fixed the behavior when the update lambda changed the lambda name

  2. [LUNA-8278] The retransmission_fallback is now properly work for lambda of agent type.

    **Note: propper work of retransmission_fallback requires adjustment of a few kernel parameters on k8s node level: **

    • fs.inotify.max_user_instances - 1280

    • fs.inotify.max_user_watches - 655360

    Lower values could cause retransmission_fallback to stop working.

Migration:

See a migration manual.