@yingyeothon/logger-slack

Slack incoming-webhook backend for the @yingyeothon/logger contract. createSlackLogWriter returns a LogWriter that batches every record onto a single promise chain posting to the webhook, plus a flush() handle so serverless handlers can await delivery before exiting. createSlackLogger wraps it in a severity-filtered Logger. When webhookUrl is missing, Slack delivery is silently skipped; library code reads no environment variables — use slackLogWriterOptionsFromEnv() if you want the classic env-based configuration.

Install

npm install @yingyeothon/logger-slack

Usage

ESM:

import {
  createSlackLogger,
  slackLogWriterOptionsFromEnv,
} from "@yingyeothon/logger-slack";

const logger = createSlackLogger({
  webhookUrl: "https://hooks.slack.com/services/T000/B000/XXX",
  channel: "#alerts",
  userName: "yyt-bot",
  severity: "warn", // default; debug/info are filtered out
  withConsole: true, // also mirror records to the console
});

logger.info("request accepted", { requestId: "r-1" });
logger.error("request failed", { error: new Error("boom") }); // Error values are serialized.

// Or take webhookUrl/channel/userName from SLACK_WEBHOOK_URL etc.:
// const logger = createSlackLogger(slackLogWriterOptionsFromEnv());

// Wait for all queued Slack posts before the process exits.
await logger.flush();

CJS:

const { createSlackLogWriter } = require("@yingyeothon/logger-slack");
const writer = createSlackLogWriter({
  webhookUrl: process.env.SLACK_WEBHOOK_URL,
});
writer.warn("retrying", { jobId: 7 });
writer.flush().then(() => process.exit(0));

Each Slack message looks like [WARN] retrying followed by a fenced JSON block containing timestamp and context (the arguments after the message; omitted when there are none).

Public API

• createSlackLogWriter(options?) — LogWriter (debug/info/warn/error, variadic) plus flush(): Promise<void>; options: webhookUrl?, channel?, userName? (default "Logger"), maxTextLength? (default 24 KiB), onDeliveryError?.
• createSlackLogger(options?) — severity-filtered Logger plus flush(); adds severity? (default "warn") and withConsole? (combine with consoleWriter) to the writer options.
• slackLogWriterOptionsFromEnv() — opt-in helper reading SLACK_WEBHOOK_URL, SLACK_CHANNEL, SLACK_USER_NAME.
• Types: SlackLogWriter, SlackLogWriterOptions, SlackLogger, SlackLoggerOptions.

Migrating from the legacy package

The package was renamed on npm: @yingyeothon/slack-logger → @yingyeothon/logger-slack, and the API was redesigned around the shared @yingyeothon/logger contract.

• Seven levels (trace…silent) became the shared four (debug, info, warn, error) with a severity filter ("none" disables everything); LogLevel, parseLogLevel, and toLogLevelName are gone.
• Call style flipped from (context, message) to message-first variadic: logger.error("failed", { requestId }).
• getLogger/useLogger/flushSlack/asYlogger were replaced by createSlackLogWriter/createSlackLogger; the adapter is unnecessary because the writer is a @yingyeothon/logger LogWriter.
• Configuration is injected via options instead of read from process.env; use slackLogWriterOptionsFromEnv() to keep the env-driven behavior. CONSOLE_LOG_LEVEL/SLACK_LOG_LEVEL have no replacement — pass severity.
• Pending Slack sends are stored per writer instead of in a module-global chain; call flush() on the writer/logger you created. Webhook failures are swallowed (report them via onDeliveryError), and the caller's context objects are no longer mutated when errors are serialized.
• A typical port: getLogger("api", "handler.ts") + logger.error({ error }, "request failed") + await logger.flushSlack() becomes createSlackLogger(slackLogWriterOptionsFromEnv()) + logger.error("request failed", { error }) + await logger.flush().