Network Sink Configuration

Logly supports sending logs over various network protocols. This guide covers HTTP, TCP, UDP, and Syslog sink configuration with complete examples.

HTTP JSON Sink

POST log messages as JSON to an HTTP endpoint.

Basic HTTP Sink

python

import json
import urllib.request
from logly import logger

def http_sink(message: str) -> None:
    payload = json.dumps({"log": message}).encode("utf-8")
    request = urllib.request.Request(
        "https://logs.example.com/ingest",
        data=payload,
        headers={"Content-Type": "application/json"},
        method="POST",
    )
    urllib.request.urlopen(request, timeout=5)

logger.add(http_sink, level="INFO", enqueue=True)

HttpJsonSink (Built-in)

python

from logly import logger
from logly.integrations.telemetry import HttpJsonSink

logger.add(
    HttpJsonSink(
        endpoint="https://logs.example.com/ingest",
        headers={"Authorization": "Bearer YOUR_TOKEN"},
    ),
    level="WARNING",
    serialize=True,
)

logger.warning("Error sent via HTTP")

Custom HTTP Configuration

python

import json
import urllib.request
from logly import logger

def http_sink(message: str) -> None:
    payload = json.dumps({
        "log": message,
        "source": "myapp",
        "timestamp": "2026-07-01T14:30:00Z",
    }).encode("utf-8")

    request = urllib.request.Request(
        "https://logs.example.com/ingest",
        data=payload,
        headers={
            "Content-Type": "application/json",
            "Authorization": "Bearer YOUR_TOKEN",
            "X-Source": "myapp",
            "X-Environment": "production",
        },
        method="POST",
    )
    urllib.request.urlopen(request, timeout=10)

logger.add(
    http_sink,
    level="WARNING",
    enqueue=True,
)

HTTP with Retry Logic

python

import json
import urllib.request
import time
from logly import logger

def http_sink_with_retry(message: str, max_retries=3) -> None:
    payload = json.dumps({"log": message}).encode("utf-8")

    for attempt in range(max_retries):
        try:
            request = urllib.request.Request(
                "https://logs.example.com/ingest",
                data=payload,
                headers={"Content-Type": "application/json"},
                method="POST",
            )
            urllib.request.urlopen(request, timeout=5)
            return
        except Exception as e:
            if attempt < max_retries - 1:
                time.sleep(2 ** attempt)
            else:
                logger.error("HTTP sink failed after {} retries: {}", max_retries, e)

logger.add(http_sink_with_retry, level="WARNING", enqueue=True)

HTTP Parameters

Parameter	Type	Description
`endpoint`	`str`	URL to POST logs to
`headers`	`dict`	HTTP headers for the request
`method`	`str`	HTTP method (default: `POST`)
`timeout`	`int`	Request timeout in seconds
`batch_size`	`int`	Number of logs to batch before sending
`batch_interval`	`float`	Time between batch sends (seconds)

TCP Sink

Send logs over a TCP socket for reliable delivery.

Basic TCP Sink

python

import socket
from logly import logger

def tcp_sink(message: str) -> None:
    with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as sock:
        sock.connect(("logserver.example.com", 9000))
        sock.sendall(message.encode("utf-8"))

logger.add(tcp_sink, level="INFO")

TCP with Delimiter

python

import socket
from logly import logger

def tcp_sink_with_delimiter(message: str) -> None:
    with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as sock:
        sock.connect(("logserver.example.com", 9000))
        # Add newline delimiter for line-based protocols
        sock.sendall(message.encode("utf-8") + b"\n")

logger.add(tcp_sink_with_delimiter, level="INFO")

Persistent TCP Connection

python

import socket
from logly import logger

class PersistentTCPSink:
    def __init__(self, host: str, port: int):
        self.host = host
        self.port = port
        self.sock = None

    def connect(self):
        self.sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        self.sock.connect((self.host, self.port))

    def __call__(self, message: str) -> None:
        try:
            if self.sock is None:
                self.connect()
            self.sock.sendall(message.encode("utf-8") + b"\n")
        except (ConnectionError, OSError):
            self.sock = None
            self.connect()
            self.sock.sendall(message.encode("utf-8") + b"\n")

logger.add(PersistentTCPSink("logserver.example.com", 9000), level="INFO")

TCP Parameters

Parameter	Type	Description
`host`	`str`	Server hostname or IP
`port`	`int`	Server port
`delimiter`	`str`	Message delimiter (default: `\n`)
`timeout`	`int`	Connection timeout in seconds

UDP Sink

Send logs over a UDP socket (fire-and-forget, low latency).

Basic UDP Sink

python

import socket
from logly import logger

def udp_sink(message: str) -> None:
    with socket.socket(socket.AF_INET, socket.SOCK_DGRAM) as sock:
        sock.sendto(
            message.encode("utf-8"),
            ("logserver.example.com", 514),
        )

logger.add(udp_sink, level="WARNING")

UDP with Structured Data

python

import socket
import json
from logly import logger

def udp_structured_sink(message: str) -> None:
    payload = json.dumps({
        "log": message,
        "source": "myapp",
    }).encode("utf-8")
    with socket.socket(socket.AF_INET, socket.SOCK_DGRAM) as sock:
        sock.sendto(payload, ("logserver.example.com", 514))

logger.add(udp_structured_sink, level="WARNING")

UDP Parameters

Parameter	Type	Description
`host`	`str`	Server hostname or IP
`port`	`int`	Server port
`max_size`	`int`	Maximum packet size (default: 65535)

WARNING

UDP is unreliable by design. Messages may be lost during network congestion. Use UDP only for non-critical logs or metrics.

Syslog Sink

Send logs to a Syslog server using the standard logging.handlers.SysLogHandler.

Local Syslog

python

import logging
from logging.handlers import SysLogHandler
from logly import logger

logger.add(SysLogHandler(address="/dev/log"), level="INFO")

Remote Syslog (UDP)

python

import logging
from logging.handlers import SysLogHandler
from logly import logger

logger.add(
    SysLogHandler(address=("logserver.example.com", 514)),
    level="WARNING",
)

Remote Syslog (TCP)

python

import logging
from logging.handlers import SysLogHandler
from logly import logger

logger.add(
    SysLogHandler(address=("logserver.example.com", 514), socktype=socket.SOCK_STREAM),
    level="WARNING",
)

Windows Syslog

python

import logging
from logging.handlers import SysLogHandler
from logly import logger

# Requires syslog-ng or similar on Windows
logger.add(
    SysLogHandler(address=("localhost", 514)),
    level="INFO",
)

Syslog Facilities

Syslog facilities categorize the source of log messages:

Facility	Description	Typical Use
`LOG_USER`	Generic user-level messages	General application logs
`LOG_LOCAL0`	Local use 0	Application-specific
`LOG_LOCAL1`	Local use 1	Application-specific
`LOG_LOCAL2`	Local use 2	Application-specific
`LOG_LOCAL3`	Local use 3	Application-specific
`LOG_LOCAL4`	Local use 4	Application-specific
`LOG_LOCAL5`	Local use 5	Application-specific
`LOG_LOCAL6`	Local use 6	Application-specific
`LOG_LOCAL7`	Local use 7	Application-specific
`LOG_DAEMON`	System daemons	Daemon processes
`LOG_AUTH`	Authentication/security	Login attempts
`LOG_CRON`	Cron/scheduled tasks	Job scheduler
`LOG_KERN`	Kernel messages	System-level
`LOG_MAIL`	Mail system	Email server
`LOG_NEWS`	Network news	Usenet
`LOG_SYSLOG`	Syslog internal	Internal messages

Syslog with Custom Facility

python

import logging
from logging.handlers import SysLogHandler
from logly import logger

# Use LOG_LOCAL0 for application logs
handler = SysLogHandler(
    address=("logserver.example.com", 514),
    facility=SysLogHandler.LOG_LOCAL0,
)

logger.add(handler, level="INFO")

Syslog Severity Mapping

Logly maps its log levels to Syslog severities:

Logly Level	Syslog Severity
`TRACE`	`DEBUG`
`DEBUG`	`DEBUG`
`INFO`	`INFO`
`NOTICE`	`NOTICE`
`SUCCESS`	`INFO`
`WARNING`	`WARNING`
`ERROR`	`ERROR`
`FAIL`	`CRITICAL`
`CRITICAL`	`CRITICAL`
`FATAL`	`EMERGENCY`

Syslog Formats

RFC 3164 (BSD Syslog)

python

import logging
from logging.handlers import SysLogHandler
from logly import logger

# RFC 3164 format (default)
logger.add(
    SysLogHandler(
        address=("logserver.example.com", 514),
        facility=SysLogHandler.LOG_LOCAL0,
    ),
    level="INFO",
)

Example output:

<14>Jul  1 14:30:45 myapp[12345]: INFO | Application started

RFC 5424 (Modern Syslog)

python

import logging
from logging.handlers import SysLogHandler
from logly import logger

# For RFC 5424, use structured data
logger.add(
    SysLogHandler(
        address=("logserver.example.com", 514),
        facility=SysLogHandler.LOG_LOCAL0,
    ),
    serialize=True,
    level="INFO",
)

Example output:

<14>1 2026-07-01T14:30:45.123Z myapp 12345 - - [app@48148 service="myapp"] Application started

Network Reliability Considerations

Reliability by Protocol

Protocol	Delivery	Ordering	Reliability	Use Case
HTTP	Guaranteed	Guaranteed	High	Cloud logging, structured data
TCP	Guaranteed	Guaranteed	High	Local network, reliable delivery
UDP	Best effort	May reorder	Low	Metrics, non-critical logs
Syslog	Depends on transport	Depends	Medium	System logs, centralized logging

Error Handling

python

import json
import urllib.request
from logly import logger

def reliable_http_sink(message: str) -> None:
    try:
        payload = json.dumps({"log": message}).encode("utf-8")
        request = urllib.request.Request(
            "https://logs.example.com/ingest",
            data=payload,
            headers={"Content-Type": "application/json"},
            method="POST",
        )
        urllib.request.urlopen(request, timeout=5)
    except Exception as e:
        # Log to stderr if network sink fails
        logger.opt(exception=True).error("HTTP sink failed")

logger.add(reliable_http_sink, level="INFO", enqueue=True)

Fallback Sinks

python

from logly import logger

def primary_sink(message: str) -> None:
    # Try primary destination
    import urllib.request
    urllib.request.urlopen("https://primary.example.com/logs", data=message.encode())

def fallback_sink(message: str) -> None:
    # Fallback to local file
    with open("fallback.log", "a") as f:
        f.write(message + "\n")

# Primary: network sink
logger.add(primary_sink, level="INFO", enqueue=True, catch=True)

# Fallback: file sink
logger.add("fallback.log", level="INFO", enqueue=True)

Reconnection Behavior

Automatic Reconnection

Network sinks should handle reconnection automatically:

python

import socket
from logly import logger

class ReconnectingTCPSink:
    def __init__(self, host: str, port: int, max_retries=3):
        self.host = host
        self.port = port
        self.max_retries = max_retries
        self.sock = None

    def connect(self):
        for attempt in range(self.max_retries):
            try:
                self.sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
                self.sock.connect((self.host, self.port))
                return
            except (ConnectionError, OSError) as e:
                if attempt < self.max_retries - 1:
                    import time
                    time.sleep(2 ** attempt)
                else:
                    raise

    def __call__(self, message: str) -> None:
        try:
            if self.sock is None:
                self.connect()
            self.sock.sendall(message.encode("utf-8") + b"\n")
        except (ConnectionError, OSError):
            self.sock = None
            self.connect()
            self.sock.sendall(message.encode("utf-8") + b"\n")

logger.add(
    ReconnectingTCPSink("logserver.example.com", 9000),
    level="INFO",
    enqueue=True,
)

Queue with Retry

python

import json
import urllib.request
import time
from logly import logger

def http_sink_with_retry(message: str, max_retries=3) -> None:
    payload = json.dumps({"log": message}).encode("utf-8")

    for attempt in range(max_retries):
        try:
            request = urllib.request.Request(
                "https://logs.example.com/ingest",
                data=payload,
                headers={"Content-Type": "application/json"},
                method="POST",
            )
            urllib.request.urlopen(request, timeout=5)
            return
        except Exception as e:
            if attempt < max_retries - 1:
                time.sleep(2 ** attempt)
            else:
                logger.error("HTTP sink failed after {} retries", max_retries)

logger.add(http_sink_with_retry, level="WARNING", enqueue=True)

Complete Examples

Production HTTP Logging

python

import json
import urllib.request
from logly import logger

logger.remove()

def http_json_sink(message: str) -> None:
    payload = json.dumps({
        "log": message,
        "source": "myapp",
        "environment": "production",
    }).encode("utf-8")

    request = urllib.request.Request(
        "https://logs.example.com/ingest",
        data=payload,
        headers={
            "Content-Type": "application/json",
            "Authorization": "Bearer YOUR_TOKEN",
            "X-Source": "myapp",
        },
        method="POST",
    )
    urllib.request.urlopen(request, timeout=5)

# HTTP sink for warnings and above
logger.add(
    http_json_sink,
    level="WARNING",
    enqueue=True,
)

# Local file for all logs
logger.add(
    "app.log",
    level="DEBUG",
    rotation="daily",
    retention="30 days",
    compression="gzip",
    enqueue=True,
)

# Console for immediate feedback
logger.add(
    "stderr",
    level="INFO",
    format="<green>{time:HH:mm:ss}</green> | <level>{level:<8}</level> | {message}",
    colorize=True,
)

logger.info("Application started")
logger.warning("Disk usage high")
logger.error("Connection failed")
logger.complete()

Multi-Protocol Logging

python

import socket
import json
import urllib.request
from logly import logger

logger.remove()

# TCP for reliable local delivery
def tcp_sink(message: str) -> None:
    with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as sock:
        sock.connect(("logserver.local", 9000))
        sock.sendall(message.encode("utf-8") + b"\n")

# UDP for metrics (fire-and-forget)
def udp_sink(message: str) -> None:
    with socket.socket(socket.AF_INET, socket.SOCK_DGRAM) as sock:
        sock.sendto(message.encode("utf-8"), ("logserver.local", 514))

# HTTP for cloud logging
def http_sink(message: str) -> None:
    payload = json.dumps({"log": message}).encode("utf-8")
    request = urllib.request.Request(
        "https://logs.example.com/ingest",
        data=payload,
        headers={"Content-Type": "application/json"},
        method="POST",
    )
    urllib.request.urlopen(request, timeout=5)

# All logs to TCP
logger.add(tcp_sink, level="DEBUG", enqueue=True)

# Warnings to HTTP
logger.add(http_sink, level="WARNING", enqueue=True)

# Metrics to UDP
logger.add(udp_sink, level="SUCCESS", enqueue=True)

# Console
logger.add("stderr", level="INFO", colorize=True)

logger.info("Debug to TCP and console")
logger.warning("Warning to all destinations")
logger.error("Error to all destinations")
logger.complete()

Syslog with Custom Facility

python

import logging
from logging.handlers import SysLogHandler
from logly import logger

logger.remove()

# Application logs to LOG_LOCAL0
app_handler = SysLogHandler(
    address=("syslog.example.com", 514),
    facility=SysLogHandler.LOG_LOCAL0,
)

# Security logs to LOG_AUTH
security_handler = SysLogHandler(
    address=("syslog.example.com", 514),
    facility=SysLogHandler.LOG_AUTH,
)

logger.add(app_handler, level="INFO")
logger.add(security_handler, level="WARNING")

# Custom level for security events
logger.level("SECURITY", no=45, color="bold red")

logger.info("Application started")
logger.warning("Configuration issue")
logger.log("SECURITY", "Unauthorized access attempt")
logger.complete()

High-Throughput Network Logging

python

import json
import urllib.request
from logly import logger
import time

logger.remove()

def batch_http_sink(message: str) -> None:
    payload = json.dumps({"log": message, "ts": time.time()}).encode("utf-8")
    request = urllib.request.Request(
        "https://logs.example.com/batch",
        data=payload,
        headers={"Content-Type": "application/json"},
        method="POST",
    )
    urllib.request.urlopen(request, timeout=10)

# High-throughput with batch processing
logger.add(
    batch_http_sink,
    level="INFO",
    enqueue=True,
)

# Process millions of events
for i in range(1000000):
    logger.info("Event {}", i)
    if i % 10000 == 0:
        time.sleep(0.01)  # Simulate processing time

logger.complete()

Network Sink Configuration ​

HTTP JSON Sink ​

Basic HTTP Sink ​

HttpJsonSink (Built-in) ​

Custom HTTP Configuration ​

HTTP with Retry Logic ​

HTTP Parameters ​

TCP Sink ​

Basic TCP Sink ​

TCP with Delimiter ​

Persistent TCP Connection ​

TCP Parameters ​

UDP Sink ​

Basic UDP Sink ​

UDP with Structured Data ​

UDP Parameters ​

Syslog Sink ​

Local Syslog ​

Remote Syslog (UDP) ​

Remote Syslog (TCP) ​

Windows Syslog ​

Syslog Facilities ​

Syslog with Custom Facility ​

Syslog Severity Mapping ​

Syslog Formats ​

RFC 3164 (BSD Syslog) ​

RFC 5424 (Modern Syslog) ​

Network Reliability Considerations ​

Reliability by Protocol ​

Error Handling ​

Fallback Sinks ​

Reconnection Behavior ​

Automatic Reconnection ​

Queue with Retry ​

Complete Examples ​

Production HTTP Logging ​

Multi-Protocol Logging ​

Syslog with Custom Facility ​

High-Throughput Network Logging ​

Network Sink Configuration

HTTP JSON Sink

Basic HTTP Sink

HttpJsonSink (Built-in)

Custom HTTP Configuration

HTTP with Retry Logic

HTTP Parameters

TCP Sink

Basic TCP Sink

TCP with Delimiter

Persistent TCP Connection

TCP Parameters

UDP Sink

Basic UDP Sink

UDP with Structured Data

UDP Parameters

Syslog Sink

Local Syslog

Remote Syslog (UDP)

Remote Syslog (TCP)

Windows Syslog

Syslog Facilities

Syslog with Custom Facility

Syslog Severity Mapping

Syslog Formats

RFC 3164 (BSD Syslog)

RFC 5424 (Modern Syslog)

Network Reliability Considerations

Reliability by Protocol

Error Handling

Fallback Sinks

Reconnection Behavior

Automatic Reconnection

Queue with Retry

Complete Examples

Production HTTP Logging

Multi-Protocol Logging

Syslog with Custom Facility

High-Throughput Network Logging