@debugr/html v3.0.0-rc.9

Html Log Handler for Debugr

This Handler will create interactive HTML dump files for each top-level task which contains at least one entry at or above a configured threshold.

Installation

npm install --save @debugr/html

Usage

import { Logger, LogLevel } from '@debugr/core';
import { HtmlHandler } from '@debugr/html';

const globalContext = {
  applicationName: 'example',
};

const logger = new Logger(globalContext, [
  new HtmlHandler({
    outputDir: './log',
  }),
]);

logger.runTask(async () => {
  logger.info('Task started.');
  
  await someWorkBeingDone();
  
  logger.info('Work has been done.');
  
  await moreBoringThingsHappeningInTheBackground();
  
  logger.info('Now it looks as though we\'re finally finished.');
});

How it works

When a task is started, the log handler will create an internal queue for that task. Each time a log entry is added, the handler will check if the entry exceeds the configured threshold, and if it does, the entire queue for the current task is marked to be written upon completion. So tasks which don't produce any log entries above the configured threshold will not create a dump file - e.g. HTTP requests which complete successfully. But if even a single entry exceeds the threshold, the entire log for the task will be written, so that you can debug step by step.

This works even with subtasks - if you e.g. separate the authentication step of a request handler into a subtask, and you log an entry during the subtask which would exceed the threshold, a dump will be created and the entire task tree for the request will be included. Notably, though, each task can have its threshold set independently. Initially each task inherits its threshold from its parent task, with the root task getting its threshold from the log handler options. During the execution of a subtask you can get the log handler instance from the Logger instance using the logger.getPlugin('html') method, and then you can use that instance to set options for the current task - change its threshold or force it to generate or to not generate a dump file regardless of whether there is an entry which exceeds the threshold.

Options

The HtmlHandler constructor accepts an options object as the first or second argument, depending on whether you wish to pass a preconfigured HtmlWriter instance. The default HtmlFileWriter simply writes the dump files into a configured directory, but you can get fancy with your own implementation, e.g. sending the files to a central API.

API

The HtmlHandler class has a couple of interesting methods mentioned above which you can use to influence whether a dump file will be generated for a particular task or subtask. Here they are:

• handler.setThreshold(threshold: LogLevel): void - This method can change the level at or above which a log entry needs to be in order for a dump file to be generated. Note that it doesn't work retroactively - an entry is compared against the threshold at the time it is logged, so if you increase the threshold after an entry already exceeded the original threshold, the dump file will still be generated; and similarly, if you decrease the threshold, previous entries above the new threshold but below the original will not cause a dump file to be generated.
• handler.markTaskForWriting(force: boolean = false): void - This method can be used to tell the handler that the current task should generate a dump file, even if no entry matches or exceeds the task's threshold. With the force argument omitted or false the method will only have effect if the task wasn't yet flagged either way - that is, if no entry has yet matched or exceeded the task's threshold and neither handler.markTaskForWriting() nor handler.markTaskIgnored() have been called. With force set to true the task will generate a dump file, no questions asked.
• handler.markTaskIgnored(force: boolean = false): void - This method does the opposite of handler.markTaskForWriting() - it tells the log handler that regardless of any entries which may reach or exceed the task's threshold the task should not generate a dump file. The force argument behaves just as it does in handler.markTaskForWriting(). Note that due to the fact that any single task which is marked for writing will cause the entire task tree to be written these two methods don't have the exact same effect - meaning that when you call handler.markTaskForWriting() (especially with force set to true), you can be sure that a dump file will be generated, whereas even if you call handler.markTaskIgnored() (even with force set to true), there's no guarantee that a different task won't cause a dump file to be generated. It seems logical and desirable from my perspective, but maybe it doesn't from yours, so I thought I'd elaborate.