GitLab pipeline to upload a file to AWS S3

Create S3 bucket

• go to S3 and click Create bucket
• enter Bucket name, for example dmitri-11102022, select AWS region, for example us-east-1, note this values, leave other options as default and click Create bucket

Create IAM programmatic account:

• go to IAM
• go to Users and click Add user and type a name, for example gitlab. Select Access key - Programmatic access
• select Attach existing policies directly, search for AmazonS3FullAccess and select it
• no tags are needed at next step
• click Create user and save the Access key ID and Secret access key from the next page

Add the keys to GitLab:

• go to Settings->CICD->Variables
• add variable AWS_ACCESS_KEY_ID and uncheck Protect variable and Mask variable
• add variable AWS_SECRET_ACCESS_KEY and uncheck Protect variable and check Mask variable - to hide this variable
• add variable AWS_DEFAULT_REGION and uncheck Protect variable and Mask variable, put the value saved from Create bucket step, for example us-east-1
• add variable AWS_S3_BUCKET with value from Create bucket step (example - dmitri-11102022) and uncheck Protect variable and Mask variable

Create .gitlab-ci.yaml in GitLab with the following contents:

deploy to s3:
    image:
        name: amazon/aws-cli
        entrypoint: [""]
    script:
        - aws --version
        - echo "Hello S3" > file.txt
	- aws s3 cp file.txt s3://$AWS_S3_BUCKET/file.txt

Commit and wait for pipeline to execute. Now you have your file on S3.

In other cases this option also could be used:

aws sync source_dir s3://bucket_name --delete