scalo

The runtime layer for Python control-plane services: config, logging, metrics, health probes, secrets and a service CLI, already wired together. Subclass one class and your service reads its config, logs in the right format for where it runs, and serves its health probes.

It is built for APIs, orchestrators, CLI tools and integration glue. For the hot path there is a Rust sister library, scalo-rs, with the same conventions and a different API.

What it gives you

• A seven-layer config cascade, from CLI flags and env vars down to YAML and defaults, that reads the same in a laptop shell and a Kubernetes pod.
• Structured logs, JSON in a container and readable text at a terminal. Credentials, tokens and keys are masked, and so are cards, IBANs, emails and phone numbers, and Australian ABN, ACN, TFN and Medicare numbers where a keyword names them. TRACE records skip the scrubber by default.
• ServiceApp: subclass it and get run, version, config-check and generate-artefacts, with config and logging started for you, and metrics too once scalo[metrics] is installed.
• /livez and /readyz on an observability port kept apart from your application's own.
• With the metrics extra, Prometheus and OpenTelemetry metrics, process and container gauges included, scraped from /metrics on the same port. Without a metrics manager that path answers 404.
• Secrets from OpenBao or Vault, AWS, GCP, Azure, and Kubernetes, one extra per backend.
• A deployment contract: return one from deployment_contract() and generate-artefacts writes your runtime Dockerfile and the contract the scalo-service Helm chart renders from.
• Kafka, CEL expressions and an HTTP client with jittered retry, each behind its own extra.

Example

uv add scalo

from scalo.cli import ServiceApp, VersionInfo
from scalo.logger import logger


class Greeter(ServiceApp):
    name = "greeter"
    env_prefix = "GREETER"

    def version_info(self) -> VersionInfo:
        return VersionInfo(self.name, "1.0.0")

    def run_service(self, config) -> None:
        logger.info("hello", target=config.get("target", "world"))


if __name__ == "__main__":
    Greeter().cli()

GREETER_TARGET=team python greeter.py run logs hello with target=team, and python greeter.py config-check validates the config without starting anything.

Extras

Pick the secrets backend you use rather than secrets, since each backend brings its own cloud SDK.

Docs and source are at https://github.com/hyperi-io/scalo-py. Requires Python 3.14 or later. Licensed Apache-2.0: https://github.com/hyperi-io/scalo-py/blob/main/LICENSE.