Honeycomb Sink

Behavior and Context

The sink initializes libhoney globally with write key and dataset, then creates an event per log call and attaches log fields plus structured extras.

Purpose

• Observability Enrichment: Route structured operational events into Honeycomb.
• Simple Event Model: One log call maps to one Honeycomb event.
• Minimal Setup: API key and dataset are sufficient.

High-Level API & Examples

Example 1: Basic Honeycomb Setup

from jazzmine.logging.sinks.honeycomb_sink import HoneycombLogger

sink = HoneycombLogger(
    sink_config={
        "type": "honeycomb",
        "api_key": "<honeycomb_key>",
        "dataset": "chat-api-logs",
    },
    logger_name="chat-api",
    level="INFO",
    use_json=True,
)

sink.initialize()
sink.info("request completed", status=200, latency_ms=23)
sink.shutdown()

Example 2: Async Path

import asyncio

async def send_async(sink):
    await sink.alog("warning", "high latency", p99_ms=710)

# alog delegates to synchronous log() in this sink.

Detailed Class Functionality

initialize()

• Imports libhoney lazily.
• Raises LoggerDependencyError if missing.
• Reads api_key and optional dataset (defaults to logger name).
• Calls libhoney.init(writekey=..., dataset=...).

log(level, msg, **kwargs)

• Creates a new Honeycomb event.
• Adds core fields: level, message, logger.
• Adds all extra kwargs as event fields.
• Sends event via event.send().

shutdown()

Closes libhoney client with libhoney.close().

alog(...)

Delegates to log(...).

Error Handling

• LoggerDependencyError when libhoney is unavailable.
• Event send exceptions are swallowed.

Remarks

• Because libhoney uses process-wide initialization, coordinate with any other Honeycomb integrations in the same runtime.
• Confirm dataset naming conventions with observability governance standards.