What's New

The last month has seen the release of a laundry list of features to help improve performance, visibility, and quality of life.

• Prepayment When you first add a payment method, you'll be asked to prepay $50 in return for $25 credit, providing you $75 in total credit that will be automatically applied toward any deployments on Salad Cloud. Once you've prepaid, you'll receive a small amount of quota (currently, this is set at 2 Container Group Deployments, 2 Recipe Deployments, with 3 replicas for each) in order to deploy a small application or conduct integration tests. Reach out to your account executive or fill out the quota increase requests forms in the portal when you are ready to increase your quota.

• HTTP Probes support You're now able to configure startup and liveness health probes using the HTTP protocol.

  

• Better instance statuses We receive many questions about how long setting up an instance should be expected to take. Now, when you deploy a container group, you will now see individual progress bars to indicate the downloading and creating processes on allocated nodes. Click on any of these instance status indicators to see additional information about what the statuses mean. Note that recipes do not yet support these more detailed status indicators. Also note that the statuses are based on estimated download and container create times given the size of your specific container image to give a better indication of when your instance will reach a running state. We're currently working on enhancing these progress indicators with real-time percentages from the nodes themselves.

  

• Datadog logging support You can now stream logs to Datadog from your containers. You'll need to define the destination during configuration of your Container Group Deployment.

  

• New Relic logging support You can now stream logs to New Relic from your containers. You'll need to define the destination during configuration of your Container Group Deployment.

  

• HTTP logging support You can now stream logs to an arbitrary external host over HTTP. You'll need to define the destination during configuration of your Container Group Deployment.

  

• Auto-start Previously in the portal, you had to monitor any new deployment and then press the start button once it had finished pulling the image in order to actually deploy to nodes. Now, you can configure your container to auto-start once pulling is completed. Auto-start is now the default behavior for new Container Groups.

  

• Faster image pulling We've increased the speed to pull an image (the 'preparing' step during deployment).

• Duplicate a container group Sometimes you may want to tweak the configuration of an existing container group, such as to increase the RAM allocation or to update to the latest version of the image. Previously in the portal, you had to create an entirely new container group from scratch. Now, you can simply duplicate an existing container group. This copies the existing configuration into a new container group configuration page, where you can adjust it before deploying.

  

• High-demand RAM labels and notification Nodes with the greatest amount of RAM are occasionally in high demand. Similar to GPU cards, you may see a 'High Demand' label on RAM cards which indicates that deployments requiring that configuration may take longer to allocate all replicas.

  

• Workload Errors in Portal If your container group deployment encounters problems, take a look at the Recent Errors tab on the container group page. It'll give you a breakdown of the last hours' errors so you have an idea of what's going on. This information is also available from an API endpoint at: https://api.salad.com/api/public/organizations/{organization_name}/projects/{project_name}/containers/{container_group_name}/errors