Basic Example

Create an instance of the logger:

import { logger, http } from '@elliemae/pui-diagnostics';

const appLogger = logger({
  transport: http('https://int.api.ellielabs.com/diagnostics/v2/logging'),
  index: 'myapp',
  team: 'my team',
  appName: 'Hello World App',
});

Log messages with different log levels:

appLogger.debug('Detailed debug info');
appLogger.info('My application log message');
appLogger.audit('User performed action');
appLogger.warn('Something unexpected');
appLogger.error('Something went wrong');
appLogger.fatal('Application crashed');

Logging Exceptions

try {
  // ...
} catch (ex) {
  appLogger.error({
    code: 'puiweb002',
    message: 'Unable to send message',
    exception: ex as Error,
  });
}

Changing Default Log Level

The default log level is INFO. Messages below this level (e.g. DEBUG) are not logged.

import { LogLevels } from '@elliemae/pui-diagnostics';

appLogger.setLogLevel(LogLevels.DEBUG);

Updating Logger Options at Runtime

appLogger.setOptions({
  environment: 'prod',
  instanceId: 'new-instance',
  customerId: 'cust-456',
  userId: 'user-789',
  appVersion: '2.0.0',
  loanId: 'loan-guid',
});

Logging Unhandled Errors

Automatically captures unhandled promise rejections and uncaught errors:

import { logUnhandledErrors } from '@elliemae/pui-diagnostics';

logUnhandledErrors(appLogger);

Logging Web Vitals

Logs Core Web Vitals (LCP, INP, CLS) metrics automatically:

import { webvitals } from '@elliemae/pui-diagnostics';

webvitals(appLogger);

Controlling Log Volume with HTTP Transport Options

The http transport accepts an optional second argument to control batching, rate limiting, and sampling. These options help reduce API load from high-volume applications.

const transport = http('https://api.ellielabs.com/diagnostics/v2/logging', {
  flushInterval: 15000, // flush every 15 seconds (default: 10000)
  maxBatchSize: 50, // flush early at 50 buffered logs (default: 50)
  maxLogsPerInterval: 200, // drop non-critical logs beyond 200 per window (default: 0 = unlimited)
  samplingRate: 0.5, // send ~50% of non-critical logs (default: 1 = all)
});

const appLogger = logger({
  transport,
  index: 'myapp',
  team: 'my team',
  appName: 'Hello World App',
});

Options Reference

Option	Type	Default	Description
flushInterval	number	10000	Milliseconds between batch flushes for non-critical logs
maxBatchSize	number	50	Flush the batch early when it reaches this size
maxLogsPerInterval	number	0	Max non-critical logs per interval window. 0 = unlimited
samplingRate	number	1	Sampling rate for non-critical logs (0–1). 1 = send all, 0.1 = send ~10%

Note: WARN, ERROR, and FATAL logs are always sent immediately and are never subject to batching, rate limiting, or sampling.

Batched HTTP delivery

Uses sendBeacon, then fetch (keepalive) on the same body if needed; oversized multi-row batches may be split and retried after failures. Missing or throwing sendBeacon falls back to fetch. Failures log with console.warn. Critical logs: sendBeacon, then fetch for one message. Prefer a moderate maxBatchSize and compact payloads.

Using the Console Transport

For development or debugging, use the console transport instead of HTTP:

import { logger, Console } from '@elliemae/pui-diagnostics';

const appLogger = logger({
  transport: Console(),
  index: 'myapp',
  team: 'my team',
  appName: 'Hello World App',
});

Mirroring All Logs to Console

Send logs to your transport and the browser console:

const appLogger = logger({
  transport: http('https://api.ellielabs.com/diagnostics/v2/logging'),
  index: 'myapp',
  team: 'my team',
  appName: 'Hello World App',
  logEverythingToConsole: true,
});

By default, only ERROR and FATAL logs are mirrored to the console.