microbootstrap 0.5.2

Package for bootstrapping new micro-services

microbootstrap assists you in creating applications with all the necessary instruments already set up.

# settings.py
from microbootstrap import LitestarSettings


class YourSettings(LitestarSettings):
    ...  # Your settings are stored here


settings = YourSettings()


# application.py
import litestar
from microbootstrap.bootstrappers.litestar import LitestarBootstrapper

from your_application.settings import settings

# Use the Litestar application!
application: litestar.Litestar = LitestarBootstrapper(settings).bootstrap()

With microbootstrap, you receive an application with lightweight built-in support for:

• sentry
• prometheus
• opentelemetry
• logging
• cors
• swagger - with additional offline version support
• health-checks

Those instruments can be bootstrapped for:

• fastapi
• litestar

Interested? Let's dive right in ⚡

Table of Contents

• Installation
• Quickstart
• Settings
• Service settings
• Instruments
  • Sentry
  • Prometheus
  • Opentelemetry
  • Logging
  • CORS
  • Swagger
  • Health checks
• Configuration
  • Instruments configuration
  • Application configuration
• Advanced

Installation

You can install the package using either pip or poetry. Also, you can specify extras during installation for concrete framework:

• fastapi
• litestar

For poetry:

$ poetry add microbootstrap -E fastapi

For pip:

$ pip install microbootstrap[fastapi]

Quickstart

To configure your application, you can use the settings object.

from microbootstrap import LitestarSettings


class YourSettings(LitestarSettings):
    # General settings
    service_name: str = "my-awesome-service"

    # Sentry settings
    sentry_dsn: str = "your-sentry-dsn"

    # Prometheus settings
    prometheus_metrics_path: str = "/my-path"

    # Opentelemetry settings
    opentelemetry_container_name: str = "your-container"
    opentelemetry_endpoint: str = "/opentelemetry-endpoint"



settings = YourSettings()

Next, use the Bootstrapper object to create an application based on your settings.

import litestar
from microbootstrap.bootstrappers.litestar import LitestarBootstrapper

application: litestar.Litestar = LitestarBootstrapper(settings).bootstrap()

This approach will provide you with an application that has all the essential instruments already set up for you.

Settings

The settings object is the core of microbootstrap.

All framework-related settings inherit from the BaseServiceSettings object. BaseServiceSettings defines parameters for the service and various instruments.

However, the number of parameters is not confined to those defined in BaseServiceSettings. You can add as many as you need.

These parameters can be sourced from your environment. By default, no prefix is added to these parameters.

Example:

class YourSettings(BaseServiceSettings):
    service_name: str = "micro-service"

    your_awesome_parameter: str = "really awesome"

    ... # Other settings here

To source your_awesome_parameter from the environment, set the environment variable named YOUR_AWESOME_PARAMETER.

If you prefer to use a prefix when sourcing parameters, set the ENVIRONMENT_PREFIX environment variable in advance.

Example:

$ export ENVIRONMENT_PREFIX=YOUR_PREFIX_

Then the settings object will attempt to source the variable named YOUR_PREFIX_YOUR_AWESOME_PARAMETER.

Service settings

Each settings object for every framework includes service parameters that can be utilized by various instruments.

You can configure them manually, or set the corresponding environment variables and let microbootstrap to source them automatically.

from microbootstrap.settings import BaseServiceSettings


class ServiceSettings(BaseServiceSettings):
    service_environment: str | None = None
    service_name: str = "micro-service"
    service_description: str = "Micro service description"
    service_version: str = "1.0.0"

    ... # Other settings here

Instruments

At present, the following instruments are supported for bootstrapping:

• sentry
• prometheus
• opentelemetry
• logging
• cors
• swagger

Let's clarify the process required to bootstrap these instruments.

Sentry

To bootstrap Sentry, you must provide at least the sentry_dsn. Additional parameters can also be supplied through the settings object.

from microbootstrap.settings import BaseServiceSettings


class YourSettings(BaseServiceSettings):
    service_environment: str | None = None

    sentry_dsn: str | None = None
    sentry_traces_sample_rate: float | None = None
    sentry_sample_rate: float = pydantic.Field(default=1.0, le=1.0, ge=0.0)
    sentry_max_breadcrumbs: int = 15
    sentry_max_value_length: int = 16384
    sentry_attach_stacktrace: bool = True
    sentry_integrations: list[Integration] = []
    sentry_additional_params: dict[str, typing.Any] = {}

    ... # Other settings here

These settings are subsequently passed to the sentry-sdk package, finalizing your Sentry integration.

Prometheus

Prometheus integration presents a challenge because the underlying libraries for FastAPI and Litestar differ significantly, making it impossible to unify them under a single interface. As a result, the Prometheus settings for FastAPI and Litestar must be configured separately.

Fastapi

To bootstrap prometheus you have to provide prometheus_metrics_path

from microbootstrap.settings import FastApiSettings


class YourFastApiSettings(FastApiSettings):
    service_name: str

    prometheus_metrics_path: str = "/metrics"
    prometheus_metrics_include_in_schema: bool = False
    prometheus_instrumentator_params: dict[str, typing.Any] = {}
    prometheus_instrument_params: dict[str, typing.Any] = {}
    prometheus_expose_params: dict[str, typing.Any] = {}

    ... # Other settings here

Parameters description:

• service_name - will be attached to metric's names, but has to be named in snake_case.
• prometheus_metrics_path - path to metrics handler.
• prometheus_metrics_include_in_schema - whether to include metrics route in OpenAPI schema.
• prometheus_instrumentator_params - will be passed to Instrumentor during initialization.
• prometheus_instrument_params - will be passed to Instrumentor.instrument(...).
• prometheus_expose_params - will be passed to Instrumentor.expose(...).

FastApi prometheus bootstrapper uses prometheus-fastapi-instrumentator that's why there are three different dict for parameters.

Fastapi

To bootstrap prometheus you have to provide prometheus_metrics_path

from microbootstrap.settings import LitestarSettings


class YourFastApiSettings(LitestarSettings):
    service_name: str

    prometheus_metrics_path: str = "/metrics"
    prometheus_additional_params: dict[str, typing.Any] = {}

    ... # Other settings here

Parameters description:

• service_name - will be attached to metric's names, there are no name restrictions.
• prometheus_metrics_path - path to metrics handler.
• prometheus_additional_params - will be passed to litestar.contrib.prometheus.PrometheusConfig.

Opentelemetry

To bootstrap Opentelemetry, you must provide several parameters:

• service_name
• service_version
• opentelemetry_endpoint
• opentelemetry_namespace
• opentelemetry_container_name

However, additional parameters can also be supplied if needed.

from microbootstrap.settings import BaseServiceSettings
from microbootstrap.instruments.opentelemetry_instrument import OpenTelemetryInstrumentor


class YourSettings(BaseServiceSettings):
    service_name: str
    service_version: str

    opentelemetry_container_name: str | None = None
    opentelemetry_endpoint: str | None = None
    opentelemetry_namespace: str | None = None
    opentelemetry_insecure: bool = True
    opentelemetry_instrumentors: list[OpenTelemetryInstrumentor] = []
    opentelemetry_exclude_urls: list[str] = []

    ... # Other settings here

Parameters description:

• service_name - will be passed to the Resource.
• service_version - will be passed to the Resource.
• opentelemetry_endpoint - will be passed to OTLPSpanExporter as endpoint.
• opentelemetry_namespace - will be passed to the Resource.
• opentelemetry_insecure - is opentelemetry connection secure.
• opentelemetry_container_name - will be passed to the Resource.
• opentelemetry_instrumentors - a list of extra instrumentors.
• opentelemetry_exclude_urls - list of ignored urls.

These settings are subsequently passed to opentelemetry, finalizing your Opentelemetry integration.

Logging

microbootstrap provides in-memory JSON logging through the use of structlog. For more information on in-memory logging, refer to MemoryHandler.

To utilize this feature, your application must be in non-debug mode, meaning service_debug should be set to False, which is the default.

import logging

from microbootstrap.settings import BaseServiceSettings


class YourSettings(BaseServiceSettings):
    service_debug: bool = False

    logging_log_level: int = logging.INFO
    logging_flush_level: int = logging.ERROR
    logging_buffer_capacity: int = 10
    logging_unset_handlers: list[str] = ["uvicorn", "uvicorn.access"]
    logging_extra_processors: list[typing.Any] = []
    logging_exclude_endpoints: list[str] = []

Parameters description:

• logging_log_level - The default log level.
• logging_flush_level - All messages will be flushed from the buffer when a log with this level appears.
• logging_buffer_capacity - The number of messages your buffer will store before being flushed.
• logging_unset_handlers - Unset logger handlers.
• logging_extra_processors - Set additional structlog processors if needed.
• logging_exclude_endpoints - Exclude logging on specific endpoints.

CORS

from microbootstrap.settings import BaseServiceSettings


class YourSettings(BaseServiceSettings):
    cors_allowed_origins: list[str] = pydantic.Field(default_factory=list)
    cors_allowed_methods: list[str] = pydantic.Field(default_factory=list)
    cors_allowed_headers: list[str] = pydantic.Field(default_factory=list)
    cors_exposed_headers: list[str] = pydantic.Field(default_factory=list)
    cors_allowed_credentials: bool = False
    cors_allowed_origin_regex: str | None = None
    cors_max_age: int = 600

Parameter descriptions:

• cors_allowed_origins - A list of origins that are permitted.
• cors_allowed_methods - A list of HTTP methods that are allowed.
• cors_allowed_headers - A list of headers that are permitted.
• cors_exposed_headers - A list of headers that are exposed via the 'Access-Control-Expose-Headers' header.
• cors_allowed_credentials - A boolean value that dictates whether or not to set the 'Access-Control-Allow-Credentials' header.
• cors_allowed_origin_regex - A regex used to match against origins.
• cors_max_age - The response caching Time-To-Live (TTL) in seconds, defaults to 600.

Swagger

from microbootstrap.settings import BaseServiceSettings


class YourSettings(BaseServiceSettings):
    service_name: str = "micro-service"
    service_description: str = "Micro service description"
    service_version: str = "1.0.0"
    service_static_path: str = "/static"

    swagger_path: str = "/docs"
    swagger_offline_docs: bool = False
    swagger_extra_params: dict[str, Any] = {}

Parameter descriptions:

• service_name - The name of the service, which will be displayed in the documentation.
• service_description - A brief description of the service, which will also be displayed in the documentation.
• service_version - The current version of the service.
• service_static_path - The path for static files in the service.
• swagger_path - The path where the documentation can be found.
• swagger_offline_docs - A boolean value that, when set to True, allows the Swagger JS bundles to be accessed offline. This is because the service starts to host via static.
• swagger_extra_params - Additional parameters to pass into the OpenAPI configuration.

Health checks

from microbootstrap.settings import BaseServiceSettings


class YourSettings(BaseServiceSettings):
    service_name: str = "micro-service"
    service_version: str = "1.0.0"

    health_checks_enabled: bool = True
    health_checks_path: str = "/health/"

Parameter descriptions:

• service_name - Will be displayed in health check response.
• service_version - Will be displayed in health check response.
• health_checks_enabled - Must be True to enable health checks.
• health_checks_path - Path for health check handler.

Configuration

While settings provide a convenient mechanism, it's not always feasible to store everything within them.

There may be cases where you need to configure a tool directly. Here's how it can be done.

Instruments configuration

To manually configure an instrument, you need to import one of the available configurations from microbootstrap:

• SentryConfig
• OpentelemetryConfig
• PrometheusConfig
• LoggingConfig
• SwaggerConfig
• CorsConfig

These configurations can then be passed into the .configure_instrument or .configure_instruments bootstrapper methods.

import litestar

from microbootstrap.bootstrappers.litestar import LitestarBootstrapper
from microbootstrap import SentryConfig, OpentelemetryConfig


application: litestar.Litestar = (
    LitestarBootstrapper(settings)
    .configure_instrument(SentryConfig(sentry_dsn="https://new-dsn"))
    .configure_instrument(OpentelemetryConfig(opentelemetry_endpoint="/new-endpoint"))
    .bootstrap()
)

Alternatively,

import litestar

from microbootstrap.bootstrappers.litestar import LitestarBootstrapper
from microbootstrap import SentryConfig, OpentelemetryConfig


application: litestar.Litestar = (
    LitestarBootstrapper(settings)
    .configure_instruments(
        SentryConfig(sentry_dsn="https://[e-mail]/0"),
        OpentelemetryConfig(opentelemetry_endpoint="/new-endpoint")
    )
    .bootstrap()
)

Application configuration

The application can be configured in a similar manner:

import litestar

from microbootstrap.config.litestar import LitestarConfig
from microbootstrap.bootstrappers.litestar import LitestarBootstrapper
from microbootstrap import SentryConfig, OpentelemetryConfig


@litestar.get("/my-handler")
async def my_handler() -> str:
    return "Ok"

application: litestar.Litestar = (
    LitestarBootstrapper(settings)
    .configure_application(LitestarConfig(route_handlers=[my_handler]))
    .bootstrap()
)

Important

When configuring parameters with simple data types such as: str, int, float, etc., these variables overwrite previous values.

Example:

from microbootstrap import LitestarSettings, SentryConfig


class YourSettings(LitestarSettings):
    sentry_dsn: str = "https://my-sentry-dsn"


application: litestar.Litestar = (
    LitestarBootstrapper(YourSettings())
    .configure_instrument(
        SentryConfig(sentry_dsn="https://my-new-configured-sentry-dsn")
    )
    .bootstrap()
)

In this example, the application will be bootstrapped with the new https://my-new-configured-sentry-dsn Sentry DSN, replacing the old one.

However, when you configure parameters with complex data types such as: list, tuple, dict, or set, they are expanded or merged.

Example:

from microbootstrap import LitestarSettings, PrometheusConfig


class YourSettings(LitestarSettings):
    prometheus_additional_params: dict[str, Any] = {"first_value": 1}


application: litestar.Litestar = (
    LitestarBootstrapper(YourSettings())
    .configure_instrument(
        PrometheusConfig(prometheus_additional_params={"second_value": 2})
    )
    .bootstrap()
)

In this case, Prometheus will receive {"first_value": 1, "second_value": 2} inside prometheus_additional_params. This is also true for list, tuple, and set.

Advanced

If you miss some instrument, you can add your own. Essentialy, Instrument is just a class with some abstractmethods. Every instrument uses some config, so that's first thing, you have to define.

from microbootstrap.instruments.base import BaseInstrumentConfig


class MyInstrumentConfig(BaseInstrumentConfig):
    your_string_parameter: str
    your_list_parameter: list

Next, you can create an instrument class that inherits from Instrument and accepts your configuration as a generic parameter.

from microbootstrap.instruments.base import Instrument


class MyInstrument(Instrument[MyInstrumentConfig]):
    instrument_name: str
    ready_condition: str

    def is_ready(self) -> bool:
        pass

    def teardown(self) -> None:
        pass

    def bootstrap(self) -> None:
        pass

    @classmethod
    def get_config_type(cls) -> type[MyInstrumentConfig]:
        return MyInstrumentConfig

Now, you can define the behavior of your instrument.

Attributes:

• instrument_name - This will be displayed in your console during bootstrap.
• ready_condition - This will be displayed in your console during bootstrap if the instrument is not ready.

Methods:

• is_ready - This defines the readiness of the instrument for bootstrapping, based on its configuration values. This is required.
• teardown - This allows for a graceful shutdown of the instrument during application shutdown. This is not required.
• bootstrap - This is the main logic of the instrument. This is not required.

Once you have the framework of the instrument, you can adapt it for any existing framework. For instance, let's adapt it for litestar.

import litestar

from microbootstrap.bootstrappers.litestar import LitestarBootstrapper

@LitestarBootstrapper.use_instrument()
class LitestarMyInstrument(MyInstrument):
    def bootstrap_before(self) -> dict[str, typing.Any]:
        pass

    def bootstrap_after(self, application: litestar.Litestar) -> dict[str, typing.Any]:
        pass

To bind the instrument to a bootstrapper, use the .use_instrument decorator.

To add extra parameters to the application, you can use:

• bootstrap_before - This adds arguments to the application configuration before creation.
• bootstrap_after - This adds arguments to the application after creation.

Afterwards, you can use your instrument during the bootstrap process.

import litestar

from microbootstrap.bootstrappers.litestar import LitestarBootstrapper
from microbootstrap import SentryConfig, OpentelemetryConfig

from your_app import MyInstrumentConfig


application: litestar.Litestar = (
    LitestarBootstrapper(settings)
    .configure_instrument(
        MyInstrumentConfig(
            your_string_parameter="very-nice-parameter",
            your_list_parameter=["very-special-list"],
        )
    )
    .bootstrap()
)

Alternatively, you can fill these parameters within your main settings object.

from microbootstrap import LitestarSettings
from microbootstrap.bootstrappers.litestar import LitestarBootstrapper

from your_app import MyInstrumentConfig


class YourSettings(LitestarSettings, MyInstrumentConfig):
    your_string_parameter: str = "very-nice-parameter"
    your_list_parameter: list = ["very-special-list"]

settings = YourSettings()

application: litestar.Litestar = LitestarBootstrapper(settings).bootstrap()