Python Logging Handler

If your service already uses Python’s built-in logging module, add SquasherHandler to forward log records to Squasher without replacing the rest of your logging setup.

Installation

pip install squasher-python

Configuration

Create a handler once, attach it to your logger, and keep the rest of your logging pipeline unchanged.

app/logging.py

import logging

from squasher import SamplingConfig, SquasherHandler

handler = SquasherHandler(
    api_key="sq_pk_your_key",
    project_id="your-project-id",
    sampling=SamplingConfig(success_rate=0.1),
)

logger = logging.getLogger("app")
logger.setLevel(logging.INFO)
logger.addHandler(handler)

Option	Type	Default	Description
`api_key`	`str`	Required	Your project ingest key
`project_id`	`str`	Required	Your Squasher project ID
`endpoint`	`str`	`https://ingest.squasher.ai`	Override the ingest base URL for local development
`sampling`	`SamplingConfig`	Keep all outcomes	Set deterministic outcome sampling

The handler sends events on a background thread, so logger.error() does not block while the HTTP request is in flight.

Usage Example

app/main.py

import logging

from squasher import SquasherHandler

handler = SquasherHandler(
    api_key="sq_pk_your_key",
    project_id="your-project-id",
)

logger = logging.getLogger("app")
logger.setLevel(logging.INFO)
logger.addHandler(handler)

logger.info("worker booted")
logger.error("payment capture failed")

handler.flush()
handler.close()

Call flush() and close() during shutdown to make sure queued log records are delivered before the process exits.

Exception Handling

Use logger.exception() inside an except block to include the exception type and traceback automatically.

try:
    raise RuntimeError("card processor timed out")
except RuntimeError:
    logger.exception("captured checkout exception")
finally:
    handler.flush()
    handler.close()

When exc_info is present, the handler sends both the exception type and the formatted stack trace alongside the log message, so the error shows up in Squasher with full debugging context.

Structured attributes

Pass typed context through extra. The handler preserves numbers and booleans, redacts sensitive attribute names, and reads trace_id and span_id for correlation.

logger.info(
    "cache lookup completed",
    extra={
        "event.name": "cache.lookup.completed",
        "cache": {"hit": True, "age_ms": 42},
        "trace_id": trace_id,
        "span_id": span_id,
    },
)

SamplingConfig keeps errors and slow records by default. A stable trace gets one sampling decision.

Agent handoff

Install squasher-python, add SquasherHandler to the existing logger with SQUASHER_API_KEY and SQUASHER_PROJECT_ID from env, avoid logging secrets, call flush or close during shutdown, and verify with one logger.error call.

Get Started

SDKs

Integrations

Migrations

Cloud Log Drains

Features

API Reference

Python Logging Handler

Installation

Configuration

Usage Example

Exception Handling

Structured attributes

Agent handoff

​Installation

​Configuration

​Usage Example

​Exception Handling

​Structured attributes

​Agent handoff

Installation

Configuration

Usage Example

Exception Handling

Structured attributes

Agent handoff