Ax
12/19/2023, 12:12 AMMarvin
12/19/2023, 12:13 AMdeployment.yaml
, it suggests that such automation is possible with GitHub Actions.
For detailed examples and configuration options, you should refer to the action page on GitHub Marketplace. Typically, GitHub Actions allows you to set environmental variables or modify files using scripting within the GitHub Action workflow. You might be able to use these features to populate the branch name dynamically in your deployment.yaml
.
If you would like more specific guidance, I recommend checking the documentation or examples provided in the GitHub Action page or in the official Prefect documentation for GitHub Actions integrations. If this doesn't fully answer your question, or if you require further assistance with GitHub Actions and Prefect, please let me know, and I can assist you further.Nate
12/19/2023, 12:13 AMdeployment.yaml
or prefect.yaml
?Ax
12/19/2023, 12:17 AMprefect.yaml
Essentially want to use a github action to populate the branch name on all deployments named staging – rather than merging them to a staging branchNate
12/19/2023, 12:45 AMEssentially want to use a github action to populate the branch name on all deployments named staginghmm can you explain this a bit more? do you mean like deployments named like
staging-*
should all have the same branch
in their pull
step?Ax
12/19/2023, 9:54 AMprefect.yaml
At the moment, everything pulls for main.
name: deployments
pull:
- prefect.deployments.steps.git_clone:
repository: <https://github.com/my_github_repo.git>
branch: main
access_token: '{{ prefect.blocks.secret.dna-infrastructure-key }}'
deployments:
- name: production
version:
tags: []
description:
entrypoint: orchestration/prefect/flows/predictor.py:predictor_flow
parameters:
bucket: mlops-sagemaker-production
model_package_group_name: production-predictor
endpoint_name: production-predictor
work_pool:
name: ecs-pool
work_queue_name:
job_variables: {}
schedule:
cron: 0 7 13 * *
timezone: Europe/London
- name: staging
version:
tags: []
description:
entrypoint: orchestration/prefect/flows/predictor.py:predictor_flow
parameters:
bucket: mlops-sagemaker-staging
model_package_group_name: staging-predictor
endpoint_name: staging-predictor
work_pool:
name: ecs-pool
work_queue_name:
job_variables: {}
schedule:
cron: 0 7 13 * *
timezone: Europe/London
Ax
12/19/2023, 9:59 AM