September#

September 1#

Changes#

  • An application load balancer automatically adds the X-Forwarded-For, X-Forwarded-Port, and X-Forwarded-Proto headers to requests. A target can use them to determine the client address, as well as the port and protocol of the original connection to the load balancer.

  • To connect to the load balancer, the client can use HTTP/2 or HTTP/1.1 (HTTP/2 support can be disabled).

API#

September 10#

Changes#

  • The Bare Metal service lets you rent dedicated physical servers without virtualization. The servers are located in the same data centers as K2 Cloud. You can connect them to the cloud and use cloud services, for example Elastic IP and Network ACL.

  • The Import/Export service supports virtual machine images with UEFI firmware. You can import these images into K2 Cloud and export them back to a local virtualization infrastructure while preserving the bootloader configuration.

  • For PaaS services, you can enable and disable automatic volume expansion when the volume fills up.

  • You can add targets to a load balancer target group by IP address (type ip).

  • For internal network load balancers, you can enable traffic balancing across availability zones. You can set this option for the load balancer as a whole or override it for an individual target group.

  • You can assign additional private IP addresses to a network interface. In particular, this makes it easier to use amazon-vpc-cni in Kubernetes clusters.

  • You can assign multiple network interfaces in the same subnet to an instance.

  • You can associate a private DNS zone with multiple VPCs, including VPCs in different regions. At most five such associations are allowed for one zone.

  • In a public DNS zone of the DNSaaS service, you can create records for the DNS names of internal load balancers.

  • Regional quotas have been added for the number of GPUs in a project and for the number of instances of one family.

  • The default limits on the number of images and volume snapshots in a project have been increased to 3,000.

API#

September 11#

Changes#

  • The beta version of web interface 3.0 is available. In this version, the compute resources and networking sections have been updated, and the IAM service interface has been redesigned.

September 23#

Changes#

  • In the Quotas section, you can view the current limits and quota usage.

  • The Monitoring tab on the PaaS service page shows monitoring metric charts. Each chart is built from aggregated metrics for several resources, rather than from individual metrics for a single resource. The set of metrics differs depending on the service type.

  • For the PostgreSQL PaaS service with the pgsql1c vendor, the fasttrun, fulleq, mchar, and plantuner extensions are installed automatically.

  • Quotas have been added for private IP addresses of a network interface, wildcard attributes of a load balancer listener, and MultiValueAnswer DNS records.

  • For the KMS service, quotas have been added for the number of aliases per key and the number of encryption keys in a project.

API#

  • The DescribeAccountAttributes method lets you get the values of the ds_multivalue_rrsets_max_num, ipv4_addresses_per_interface, kms_aliases_per_key_max_num, kms_keys_max_num, and listener_wildcard_attributes_max_num quotas.

September 29#

Changes#

  • NAT functionality has been separated from the internet gateway and implemented as a separate resource — NAT gateway (NAT Gateway, NGW). A NAT gateway provides internet access for instances that do not have their own Elastic IP address.

API#

  • You can manage NAT gateways using the EC2 API.