Quickstart

Get from a new account to a successful job in your own AWS infrastructure. This guide has four stages:

1. Connect your AWS account
2. Create a network
3. Provision a cluster
4. Run your first job

Budget 10 to 15 minutes of hands-on setup, plus 15 to 25 minutes of background provisioning in AWS.

Before you begin

You'll need:

• A LakeSail account. If you don't have one yet, sign up first.
• An AWS account where you can deploy a CloudFormation stack (i.e. permission to create IAM roles).
• The AWS region you want to run workloads in.

1. Connect your AWS account

LakeSail provisions clusters inside your own AWS account using a cross-account IAM role deployed via CloudFormation. LakeSail does not store long-lived AWS access keys. It stores the role ARN, assumes the role on demand, and uses temporary credentials issued by AWS. Before approving the connection, your security team can review what the role grants, and what LakeSail can and cannot do, in Security & IAM. For the standalone connection guide, see Connect an AWS Account.

1. In the sidebar, open Settings → Cloud Accounts.
2. Click Add Account. The Add AWS Account modal opens.
3. Click Create Stack on AWS and sign in to the target AWS account if prompted.
4. In AWS CloudFormation, review and create the pre-filled stack.
5. Wait for CloudFormation to finish creating the stack. LakeSail verifies the connection automatically; you can close the modal while CloudFormation runs and return later.
6. When the modal reports AWS account connected, optionally enter an Account alias.
7. Click Save & Setup Network to continue straight to the next step.

Prefer to deploy manually?

After clicking Create Stack on AWS, return to the LakeSail modal, expand Advanced, and click Download Template. Apply it through the AWS Console, CLI, or an IaC tool. When CloudFormation finishes creating the stack, paste its role ARN into IAM Role ARN and click Verify Manually.

2. Create a network

A network is a VPC that LakeSail provisions in your connected AWS account. Clusters run inside this VPC.

1. Open Settings → Cloud Networks (or continue from step 1's Save & Setup Network).
2. Click Add Network.
3. Fill in:
   • Name, e.g. production-network.
   • Cloud Account: the one you connected in step 1.
   • Region: where the VPC should live.
   • IPv4 CIDR Range, e.g. 10.0.0.0/16. Must not overlap the LakeSail platform VPC CIDR or any VPC you intend to peer with.
4. Click Create Network. The network moves through Pending → Provisioning → Deployed as LakeSail deploys the VPC, subnets, and security groups (typically a few minutes).

LakeSail also creates the network's workspace bucket and begins provisioning its default AWS Glue catalog.

When provisioning completes, click Create Cluster to continue.

3. Provision a cluster

With an account and network in place, you can create your first cluster.

1. Open Compute → Clusters.
2. Click Create Cluster. The Create External Cluster modal opens.
3. Fill in:
   • Network: the one you just created. The cluster inherits its cloud account and region.
   • Cluster Name, e.g. my-first-cluster.
   • Cluster tier: select a tier using the guidance in the picker. The details show its management-node configuration and estimated baseline AWS cost.
   • Allowed IPs List: leave empty for private-only access. You can add CIDRs later if you need to reach the cluster endpoint from outside the VPC.
   • Collect container metrics: leave off unless you need per-pod and per-container metrics in AWS. Container logs are collected either way.
4. Click Create Cluster.

LakeSail walks through the provisioning stages (preparing, deploying to AWS, and waiting for system readiness) with a live progress bar. Provisioning typically takes 15 to 25 minutes, and the progress bar may remain on a single stage for an extended period. When you see Cluster Ready, the cluster can run workloads.

4. Run your first job

Your cluster is ready for work. You'll create a compute profile, then a SQL job that uses it, and run the job. You do not need to load data first.

1. Open Jobs in the sidebar and click Create Job.

2. Enter a Name, e.g. hello-lakesail, and pick a Team.

3. Choose SQL as the job type and Snapshot as the source, then paste:

   SELECT 'LakeSail is ready' AS status;

4. Under Compute profile, click New profile. Enter a name, select your cluster, keep Standalone execution mode, pick a driver instance type, and click Create compute profile. Select it for the job.

5. Click Save, then Run Job on the job's detail page.

A new job run moves through the run lifecycle and ends at Succeeded. Click into the run to see its timeline, job version, and timing, then Browse results to open the query output.

Next steps

To query your own tables, connect their catalog. If the S3 files are outside the network workspace bucket, make the data available to LakeSail. Then build on the job you just ran, or choose another way to run work:

• Schedule the job so it runs automatically, and set up notifications so you hear about failures.
• Save the SQL as a query when more than one job should share it.
• Create a session to connect an external Spark Connect client.
• Create a notebook to work interactively in the console.

When you are ready to collaborate, invite teammates to your organization.

Before the trial ends, set up billing through Stripe or AWS Marketplace to keep workloads running.

API reference

The infrastructure and workload steps are also available through the API:

• Cloud accounts: connect AWS, manage CloudFormation templates.
• Networks: provision and manage VPCs.
• Clusters: create, resize, destroy.