Automation and observability app

This reference is shaped like a scheduled automation command: it supports idempotent dry runs, structured output, Rich terminal presentation, JSON logs, and an optional OpenTelemetry lifecycle span. Integrations remain best-effort; missing exporters cannot change the command's exit status.

Install

python -m pip install "base-cli[rich,telemetry]"
python -m pip install ".[observability]"
base-automation --help

The first command installs framework integrations; the second installs this example. In a normal project, pin both sets of dependencies in one lock file.

Configuration

Use --environment, --config, and --log-file for lifecycle configuration. --dry-run is the safety gate for scheduled changes, and --keep-temp is useful while investigating a failed run. Real products should map a validated consumer profile into ctx.config rather than reading untrusted environment variables directly in a command.

Output and errors

base-automation --quiet --target database --format json
base-automation --quiet --dry-run --target database --format json

Both commands emit one deterministic record; the dry-run action is would-reconcile. Human text output uses Rich when installed and falls back to the built-in renderer. Invalid options return a non-zero status and keep diagnostics in the redacted log.

Tests

python -m pip install "base-cli[dev,rich,telemetry]"
python -m pytest tests

Tests cover the safety gate and verify that telemetry remains optional.

Completion

_BASE_AUTOMATION_COMPLETE=bash_source base-automation
_BASE_AUTOMATION_COMPLETE=zsh_source base-automation
_BASE_AUTOMATION_COMPLETE=fish_source base-automation

Install generated scripts through your shell's usual completion directory.

Release guidance

Publish a wheel built from a clean, locked environment, run the command against a staging target with --dry-run, and verify JSON output and log schemas before promoting. Keep OpenTelemetry exporters and Rich as optional extras so minimal automation images stay small.

Operational troubleshooting

• Add --debug --log-file /tmp/base-automation.log to collect diagnostics.
• If spans are absent, verify the OpenTelemetry SDK/provider is configured; the command intentionally treats an absent exporter as a no-op.
• Start every incident with --dry-run and compare the JSON record to the expected state before permitting writes.
• Include the run ID and redacted logs in support requests, never tokens.