Manage your Uptrack uptime monitors from your shell or CI. A small, dependency-free static binary (written in Rust) — a thin client over the public v2 API that does exactly what the dashboard does, nothing more.

# macOS / Linux — downloads the right binary for your OS/arch
curl -fsSL https://raw.githubusercontent.com/Uptrack-App/uptrack-cli/main/install.sh | sh

Or grab a binary from the latest release and put it on your PATH:

Verify with SHA256SUMS from the release.

export UPTRACK_API_KEY=utk_…                      # Settings → API keys in the dashboard
export UPTRACK_API_URL=https://api.uptrack.app    # optional, this is the default

uptrack list                       # status, type, name, url, 30-day uptime
uptrack status                     # up/down summary
uptrack create --url https://example.com [--name "Home" --type http --interval 60]
uptrack delete <name|id> [--yes]   # resolves by name or id; --yes to confirm
uptrack export -o uptrack.yaml     # dump current monitors to YAML
uptrack apply  -f uptrack.yaml     # create/update monitors + channels to match the file
  --dry-run                        # print the plan, change nothing
  --prune --yes                    # also delete resources not in the file
uptrack diff   -f uptrack.yaml     # show drift; exits non-zero if any (for CI gating)

Keep your monitors in a version-controlled YAML file and apply them from CI — review changes in a pull request, sync on merge.

monitors:
  - name: Homepage
    url: https://example.com
    type: http
    interval: 60
  - name: Login keyword
    url: https://example.com/login
    type: keyword
    keyword: "Sign in"

Monitors are matched by name (case-insensitive): a name in the file that doesn't exist yet is created; one that exists is updated if its url / type / interval / keyword differ; --prune removes monitors absent from the file. type defaults to http; interval (seconds) and keyword are optional.

Channels live alongside monitors and reconcile the same way (by name). Put secret-bearing fields behind ${ENV_VAR} references so the file is safe to commit — they're resolved at apply time, and an unset variable fails the run before any change is made:

alert_channels:
  - name: Ops Slack
    type: slack
    config:
      webhook_url: ${SLACK_WEBHOOK}   # resolved from the environment at apply time

export intentionally dumps monitors only — it never writes resolved channel secrets to disk.

uptrack diff -f uptrack.yaml prints what apply would change and exits non-zero when anything differs — run it on pull requests so a stale config fails the check.

- uses: actions/checkout@v4
- uses: Uptrack-App/uptrack-cli@v1
  with:
    command: apply              # or: diff
    config: uptrack.yaml
    # args: '--prune --yes'     # optional flags
    api-key: ${{ secrets.UPTRACK_API_KEY }}

A full pull-request-diff / push-apply workflow is in examples/github-workflow.yml.

MIT