This is a GitHub Action to be used in workflows at Flipp.
The current major version tag is v0. When using actions in this repo you should generally use this tag. The repo will continue adding patch versions v0.x and advancing the v0 tag to the latest patch version for all non-breaking changes.
This action should be included in every job in every workflow. It provides tools that should be useful across all workflows, particularly in the pre/post steps. Having all this in one action should reduce noise in your config files and allow you to get new updates "for free" in the future.
A sample configuration may look like this:
- name: Flipp global
uses: wishabi/github-actions@v0
timeout-minutes: 10
env:
SLACK_BOT_TOKEN: "${{ secrets.SLACK_BOT_TOKEN }}"
with:
slack_channel: "${{env.SLACK_CHANNEL }}"
job_status: "${{ job.status }}"
enable_ssh: trueIt doesn't matter where in the job this step appears, as only the pre/post steps are used.
GitHub Actions does not have a native Slack integration, so this action will send a notification when your job fails. By default, only failed jobs will get a notification.
Source code is based on github-action-slack-notify-build.
If you need to debug your failed jobs, you can get a free SSH session to get into the box and root around. The SSH connection info will be printed out in the "Post-Flipp Global" step so you can connect. You need to be connected to the VPN in order to SSH in.
Note that only the person who kicked off the workflow will have access. In normal cases this will be the committer. You can re-run the job manually to give access to yourself if you are not the committer.
By default SSH sessions are disabled because they will cause your jobs to look like they're hanging for 10 minutes after they've failed. You can set the enable_ssh parameter to enable them.
Source code has been copied from action-upterm.
| Configuration | Required? | Description |
|---|---|---|
| job_status | Yes | The value for this should always be "${{ job.status }}". Unfortunately GitHub Actions doesn't have a way for the action itself to figure out if the job failed. |
| slack_channel | No | The Slack channel to send updates to. Recommendation is to set a top-level environment variable (SLACK_CHANNEL) and pass it in to all calls to this action to avoid having to find/replace things later on. If this value is not provided, Slack notifications won't happen. |
| slack_always | No | If set, both success and failure notifications will be sent to Slack. |
| enable_ssh | No | If set, SSH sessions will be created on failure. |
| ssh_always | No | If set, SSH sessions will always be created, even on success. |
Please make sure to run npm run prepare before committing your files! You should probably add this to .git/hooks/pre-commit.
After creating a new minor patch version (ie v0.x), remember to update the major version tag to point to the latest patch.