GitHub Action
Careful: This package is coming soon and not published yet, so the install command below does not work today. Until it is, use the plain HTTP request from the quickstart.
One step in a workflow sends an alert to your team's phones and inbox.
Careful: Store the API key as a repository secret. Never write it into the workflow file.
Notify on a failed deploy
Add this step after your deploy. if: always() makes GitHub run it after a failed step, and on: failure makes the action send only when the job failed.
yaml
- name: Tell the team if the deploy failed
if: always() # run even after a failed step; the action decides with `on`
uses: srishrachamalla7/pingloom-notify@v1
with:
api-key: ${{ secrets.PINGLOOM_API_KEY }}
base-url: ${{ vars.PINGLOOM_BASE_URL }}
on: failure
channel: deploys
title: "Deploy failed: ${{ github.repository }}"Inputs and outputs
| Input | What it does |
|---|---|
| api-key | Required. The project API key, from a repository secret. |
| base-url | PingLoom API origin. Falls back to the PINGLOOM_BASE_URL environment variable. |
| title, body | Up to 120 and 2,000 characters. Defaults: a summary of the workflow result, and a link to the run. |
| channel | Channel slug. Default: the project's general channel. |
| priority | low, normal or high. Default: high after a failure, otherwise normal. |
| on | always (default), failure or success, compared with the job status so far. |
| idempotency-key | Default: one key per workflow run attempt. A re-run sends again; retries inside the action never duplicate. |
| fail-on-error | true fails the step if the send fails. Default false: a warning, so an alerting problem does not turn a green deploy red. |
Outputs: event-id and sent (true only when PingLoom accepted the event). The action retries rate limits and server errors with the same idempotency key, never retries a used-up quota, and masks the key in the log.