I was trying out this codelab code snippet by Google. I wanted to alter it in a way so that the Vision API metadata would be stored in a relational MySQL using the CloudSQL service - following their examples on how to connect Cloud Functions with CloudSQL.
The code I ended up deploys but upon triggering the function (by uploading a new image) I get a vague 'connection error' in the logs with no more information. This is my code at this moment:
const vision = require('#google-cloud/vision');
const Storage = require('#google-cloud/storage');
const client = new vision.ImageAnnotatorClient();
const winston = require('winston');
const {LoggingWinston} = require('#google-cloud/logging-winston');
const loggingWinston = new LoggingWinston();
const logger = winston.createLogger({
level: 'info',
transports: [new winston.transports.Console(), loggingWinston],
});
const createUnixSocketPool = async (config) => {
const dbSocketPath = "/cloudsql"
return await mysql.createPool({
user: 'root',
password: 'mypassword',
database: 'mydatabase',
socketPath: `${dbSocketPath}/cs-03-282615:europe-west1:mydatabase`,
...config
});
}
const createPool = async () => {
const config = {
connectionLimit: 5,
connectTimeout: 10000,
acquireTimeout: 10000,
waitForConnections: true,
queueLimit: 0,
}
return await createUnixSocketPool(config);
};
let pool;
const poolPromise = createPool()
.then(async (pool) => {
return pool;
})
.catch((err) => {
logger.error(err);
process.exit(1)
});
exports.vision_analysis = async (event, context, pool) => {
console.log(`Event: ${JSON.stringify(event)}`);
const filename = event.name;
const filebucket = event.bucket;
console.log(`New picture uploaded ${filename} in ${filebucket}`);
const request = {
image: { source: { imageUri: `gs://${filebucket}/${filename}` } },
features: [
{ type: 'LABEL_DETECTION' },
{ type: 'IMAGE_PROPERTIES' },
{ type: 'SAFE_SEARCH_DETECTION' }
]
};
// invoking the Vision API
const [response] = await client.annotateImage(request);
console.log(`Raw vision output for: ${filename}: ${JSON.stringify(response)}`);
if (response.error === null) {
// listing the labels found in the picture
const labels = response.labelAnnotations
.sort((ann1, ann2) => ann2.score - ann1.score)
.map(ann => ann.description)
console.log(`Labels: ${labels.join(', ')}`);
// retrieving the dominant color of the picture
const color = response.imagePropertiesAnnotation.dominantColors.colors
.sort((c1, c2) => c2.score - c1.score)[0].color;
const colorHex = decColorToHex(color.red, color.green, color.blue);
console.log(`Colors: ${colorHex}`);
// determining if the picture is safe to show
const safeSearch = response.safeSearchAnnotation;
const isSafe = ["adult", "spoof", "medical", "violence", "racy"].every(k =>
!['LIKELY', 'VERY_LIKELY'].includes(safeSearch[k]));
console.log(`Safe? ${isSafe}`);
if (isSafe) {
const pool = await poolPromise();
const stmt = 'INSERT INTO sc_03_metadata_schema (labels, color, created) VALUES (?, ?, ?)';
await pool.query(stmt, [labels, colorHex, NOW()]);
console.log("Stored metadata in CloudSQL");
}
} else {
throw new Error(`Vision API error: code ${response.error.code}, message: "${response.error.message}"`);
}
};
function decColorToHex(r, g, b) {
return '#' + Number(r).toString(16).padStart(2, '0') +
Number(g).toString(16).padStart(2, '0') +
Number(b).toString(16).padStart(2, '0');
}
The full error is as such:
Error Logs
I understand you wish to connect to your Cloud SQL MySQL instance via a Cloud Function. From referencing the public documentation Connecting from Cloud Functions to Cloud SQL it seems that “process.env.DB_SOCKET_PATH || “ was not included when instantiating your dbSocketPath.
In regards to the vague error message, this is a known issue, one of which is currently being investigated by the Cloud Functions specialists. I would suggest that you take a look at the Public Tracker and click the star icon to subscribe to the issue for further updates and click the bell icon to be notified of updates via email. We do not have an ETA on the resolution of this issue at this time.
I'm have existing code that uses express-pino-logger. This words great with our ELK stack setup, but is pretty unfortunate (logs minified JSON) when running locally.
I'd like to use pino-pretty to make local use not a pain.
There is an alternative in pino-pretty-express that solves the problem, but uses its own pretty formatter. I'd like to use the standard packages from pinojs if I could.
Here's what I have so far:
// with just pino-pretty installed, pino works out of the box
const pino = require('pino')
const logger = pino({
prettyPrint: true
})
logger.info('hi') // prints pretty
And:
// adding this option to express-pino-logger, doesn't work
const pino = require('express-pino-logger')
const logger = pino({
prettyPrint: true
})
logger.info('hi') // does NOT print pretty
I've solved my own problem I guess.
The key lies in the very last example on the express-pino-logger page:
'use strict'
const pino = require('pino')()
const expressPino = require('express-pino-logger')({
logger: pino
})
Here's my solution:
// use pino-pretty and express-pino-logger together
const basicPino = require('pino')
const basicPinoLogger = basicPino({ prettyPrint: true })
const expressPino = require('express-pino-logger')({
logger: basicPinoLogger
})
const logger = expressPino.logger
logger.info('hi') // prints pretty
const pino = require('pino');
const expressPino = require('express-pino-logger');
const logger = pino({ level: process.env.LOG_LEVEL || 'info', prettyPrint: true });
const expressLogger = expressPino({ logger });
I suppose the express pino logger is explicitly being used to log unique request ids. There is an alternative approach using the http-context and node-uuid (although I am unsure about the benchmark results)
You can create a unique uuid before the call of each request like this
const uuid = require('node-uuid');
const httpContext = require('express-http-context');
....
app.use(httpContext.middleware);
app.use((req, res, next) => {
httpContext.set('reqId', uuid.v4());
next();
});
.. you can obtain the requestId anywhere from the httpContext. No need of relying on the req object to be passed
Example usage in a custom implementation of PinoLogger service
public infoLogService (fileName): pino.Logger {
return pino({
level: 'info',
name: this.appService.getApp_name(),
messageKey: 'feedback-Logs',
base: {pid: process.pid, hostname: os.hostname,
timestamp: this.getTimeStamp(),
appName: this.appService.getApp_name(),
fileName: fileName,
request_id: isNullOrUndefined(httpContext.get('reqId')) ? 'Not an actual request ' : httpContext.get('reqId')
},
enabled: true,
useLevelLabels: true,
});
}
For every http calls, we get a UUID. for other type of loggings like logger invocation before the app starts to find if the connection to database happened successfully, there will be no uuid and hence passed a default message
Can I configure console.log so that the logs are written on a file instead of being printed in the console?
You could also just overload the default console.log function:
var fs = require('fs');
var util = require('util');
var log_file = fs.createWriteStream(__dirname + '/debug.log', {flags : 'w'});
var log_stdout = process.stdout;
console.log = function(d) { //
log_file.write(util.format(d) + '\n');
log_stdout.write(util.format(d) + '\n');
};
Above example will log to debug.log and stdout.
Edit: See multiparameter version by Clément also on this page.
Update 2013 - This was written around Node v0.2 and v0.4; There are much better utilites now around logging. I highly recommend Winston
Update Late 2013 - We still use winston, but now with a logger library to wrap the functionality around logging of custom objects and formatting. Here is a sample of our logger.js https://gist.github.com/rtgibbons/7354879
Should be as simple as this.
var access = fs.createWriteStream(dir + '/node.access.log', { flags: 'a' })
, error = fs.createWriteStream(dir + '/node.error.log', { flags: 'a' });
// redirect stdout / stderr
proc.stdout.pipe(access);
proc.stderr.pipe(error);
If you are looking for something in production winston is probably the best choice.
If you just want to do dev stuff quickly, output directly to a file (I think this works only for *nix systems):
nohup node simple-server.js > output.log &
I often use many arguments to console.log() and console.error(), so my solution would be:
var fs = require('fs');
var util = require('util');
var logFile = fs.createWriteStream('log.txt', { flags: 'a' });
// Or 'w' to truncate the file every time the process starts.
var logStdout = process.stdout;
console.log = function () {
logFile.write(util.format.apply(null, arguments) + '\n');
logStdout.write(util.format.apply(null, arguments) + '\n');
}
console.error = console.log;
Winston is a very-popular npm-module used for logging.
Here is a how-to.
Install winston in your project as:
npm install winston --save
Here's a configuration ready to use out-of-box that I use frequently in my projects as logger.js under utils.
/**
* Configurations of logger.
*/
const winston = require('winston');
const winstonRotator = require('winston-daily-rotate-file');
const consoleConfig = [
new winston.transports.Console({
'colorize': true
})
];
const createLogger = new winston.Logger({
'transports': consoleConfig
});
const successLogger = createLogger;
successLogger.add(winstonRotator, {
'name': 'access-file',
'level': 'info',
'filename': './logs/access.log',
'json': false,
'datePattern': 'yyyy-MM-dd-',
'prepend': true
});
const errorLogger = createLogger;
errorLogger.add(winstonRotator, {
'name': 'error-file',
'level': 'error',
'filename': './logs/error.log',
'json': false,
'datePattern': 'yyyy-MM-dd-',
'prepend': true
});
module.exports = {
'successlog': successLogger,
'errorlog': errorLogger
};
And then simply import wherever required as this:
const errorLog = require('../util/logger').errorlog;
const successlog = require('../util/logger').successlog;
Then you can log the success as:
successlog.info(`Success Message and variables: ${variable}`);
and Errors as:
errorlog.error(`Error Message : ${error}`);
It also logs all the success-logs and error-logs in a file under logs directory date-wise as you can see here.
const fs = require("fs");
const {keys} = Object;
const {Console} = console;
/**
* Redirect console to a file. Call without path or with false-y
* value to restore original behavior.
* #param {string} [path]
*/
function file(path) {
const con = path ? new Console(fs.createWriteStream(path)) : null;
keys(Console.prototype).forEach(key => {
if (path) {
this[key] = (...args) => con[key](...args);
} else {
delete this[key];
}
});
};
// patch global console object and export
module.exports = console.file = file;
To use it, do something like:
require("./console-file");
console.file("/path/to.log");
console.log("write to file!");
console.error("also write to file!");
console.file(); // go back to writing to stdout
For simple cases, we could redirect the Standard Out (STDOUT) and Standard Error (STDERR) streams directly to a file(say, test.log) using '>' and '2>&1'
Example:
// test.js
(function() {
// Below outputs are sent to Standard Out (STDOUT) stream
console.log("Hello Log");
console.info("Hello Info");
// Below outputs are sent to Standard Error (STDERR) stream
console.error("Hello Error");
console.warn("Hello Warning");
})();
node test.js > test.log 2>&1
As per the POSIX standard, 'input', 'output' and 'error' streams are identified by the positive integer file descriptors (0, 1, 2). i.e., stdin is 0, stdout is 1, and stderr is 2.
Step 1: '2>&1' will redirect from 2 (stderr) to 1 (stdout)
Step 2: '>' will redirect from 1 (stdout) to file (test.log)
If this is for an application, you're probably better off using a logging module. It'll give you more flexibility. Some suggestions.
winston https://github.com/winstonjs/winston
log4js https://github.com/nomiddlename/log4js-node
Straight from nodejs's API docs on Console
const output = fs.createWriteStream('./stdout.log');
const errorOutput = fs.createWriteStream('./stderr.log');
// custom simple logger
const logger = new Console(output, errorOutput);
// use it like console
const count = 5;
logger.log('count: %d', count);
// in stdout.log: count 5
Another solution not mentioned yet is by hooking the Writable streams in process.stdout and process.stderr. This way you don't need to override all the console functions that output to stdout and stderr. This implementation redirects both stdout and stderr to a log file:
var log_file = require('fs').createWriteStream(__dirname + '/log.txt', {flags : 'w'})
function hook_stream(stream, callback) {
var old_write = stream.write
stream.write = (function(write) {
return function(string, encoding, fd) {
write.apply(stream, arguments) // comments this line if you don't want output in the console
callback(string, encoding, fd)
}
})(stream.write)
return function() {
stream.write = old_write
}
}
console.log('a')
console.error('b')
var unhook_stdout = hook_stream(process.stdout, function(string, encoding, fd) {
log_file.write(string, encoding)
})
var unhook_stderr = hook_stream(process.stderr, function(string, encoding, fd) {
log_file.write(string, encoding)
})
console.log('c')
console.error('d')
unhook_stdout()
unhook_stderr()
console.log('e')
console.error('f')
It should print in the console
a
b
c
d
e
f
and in the log file:
c
d
For more info, check this gist.
Overwriting console.log is the way to go. But for it to work in required modules, you also need to export it.
module.exports = console;
To save yourself the trouble of writing log files, rotating and stuff, you might consider using a simple logger module like winston:
// Include the logger module
var winston = require('winston');
// Set up log file. (you can also define size, rotation etc.)
winston.add(winston.transports.File, { filename: 'somefile.log' });
// Overwrite some of the build-in console functions
console.error = winston.error;
console.log = winston.info;
console.info = winston.info;
console.debug = winston.debug;
console.warn = winston.warn;
module.exports = console;
If you're using linux, you can also use output redirection. Not sure about Windows.
node server.js >> file.log 2>> file.log
>> file.log to redirect stdout to the file
2>> file.log to redirect stderr to the file
others use the shorthand &>> for both stdout and stderr but it's not accepted by both my mac and ubuntu :(
extra: > overwrites, while >> appends.
By the way, regarding NodeJS loggers, I use pino + pino-pretty logger
METHOD STDOUT AND STDERR
This approach can help you (I use something similar in my projects) and works for all methods including console.log, console.warn, console.error, console.info
This method write the bytes written in stdout and stderr to file. Is better than changing console.log, console.warn, console.error, console.info methods, because output will be exact the same as this methods output
var fs= require("fs")
var os= require("os")
var HOME= os.homedir()
var stdout_r = fs.createWriteStream(HOME + '/node.stdout.log', { flags: 'a' })
var stderr_r = fs.createWriteStream(HOME + '/node.stderr.log', { flags: 'a' })
var attachToLog= function(std, std_new){
var originalwrite= std.write
std.write= function(data,enc){
try{
var d= data
if(!Buffer.isBuffer(d))
d= Buffer.from(data, (typeof enc === 'string') ? enc : "utf8")
std_new.write.apply(std_new, d)
}catch(e){}
return originalwrite.apply(std, arguments)
}
}
attachToLog(process.stdout, stdout_r)
attachToLog(process.stderr, stderr_r)
// recommended catch error on stdout_r and stderr_r
// stdout_r.on("error", yourfunction)
// stderr_r.on("error", yourfunction)
Adding to the answer above, a lit bit of an expansion to the short and efficient code overriding console.log. Minor additions: set filename with date, wrapper function, also do the original console.logging to keep the console active with the info.
Usage: in the beginning of your code, run setConsoleLogToFile([FILENAME]).
const fs = require("fs"),
util = require('util');
const getPrettyDate = ()=> new Date().toString().replace(":","-").replace(/00\s\(.*\)/, "").replace(` ${new Date().getFullYear()}`, ",").replace(/:\d\d\s/, " ");
module.exports.getPrettyDate = getPrettyDate;
module.exports.setConsoleLogToFile = (filename) => {
const log_file = fs.createWriteStream(`${__dirname}/${filename} - ${getPrettyDate()}.log`, { flags: 'w' }),
log_stdout = process.stdout;
const origConsole = console.log;
console.log = (d) => {
origConsole(d);
log_file.write(util.format(d) + '\n');
log_stdout.write(util.format(d) + '\n');
};
}
Most logger is overkill and does not support the build in console.log correctly. Hence I create console-log-to-file:
import { consoleLogToFile } from "console-log-to-file";
// or `const { consoleLogToFile } = require("console-log-to-file/dist/index.cjs.js")`
consoleLogToFile({
logFilePath: "/log/default.log",
});
// all of your console.log/warn/error/info will work as it does and save to file now.
If you are looking for a solution without modifying any code, here is a simple solution.
It requires pm2, just add it to your node modules and start you app with
pm2 start server.js
And you are done, console.logs are now automatically registered under home/.pm2/logs/server-out.log.
Improve on Andres Riofrio , to handle any number of arguments
var fs = require('fs');
var util = require('util');
var log_file = fs.createWriteStream(__dirname + '/debug.log', {flags : 'w'});
var log_stdout = process.stdout;
console.log = function(...args) {
var output = args.join(' ');
log_file.write(util.format(output) + '\r\n');
log_stdout.write(util.format(output) + '\r\n');
};
You can now use Caterpillar which is a streams based logging system, allowing you to log to it, then pipe the output off to different transforms and locations.
Outputting to a file is as easy as:
var logger = new (require('./').Logger)();
logger.pipe(require('fs').createWriteStream('./debug.log'));
logger.log('your log message');
Complete example on the Caterpillar Website
You can also have a look at this npm module:
https://www.npmjs.com/package/noogger
noogger
simple and straight forward...
For future users. #keshavDulal answer doesn't work for latest version. And I couldn't find a proper fix for the issues that are reporting in the latest version 3.3.3.
Anyway I finally fixed it after researching a bit. Here is the solution for winston version 3.3.3
Install winston and winston-daily-rotate-file
npm install winston
npm install winston-daily-rotate-file
Create a new file utils/logger.js
const winston = require('winston');
const winstonRotator = require('winston-daily-rotate-file');
var logger = new winston.createLogger({
transports: [
new (winston.transports.DailyRotateFile)({
name: 'access-file',
level: 'info',
filename: './logs/access.log',
json: false,
datePattern: 'yyyy-MM-DD',
prepend: true,
maxFiles: 10
}),
new (winston.transports.DailyRotateFile)({
name: 'error-file',
level: 'error',
filename: './logs/error.log',
json: false,
datePattern: 'yyyy-MM-DD',
prepend: true,
maxFiles: 10
})
]
});
module.exports = {
logger
};
Then in any file where you want to use logging import the module like
const logger = require('./utils/logger').logger;
Use logger like the following:
logger.info('Info service started');
logger.error('Service crashed');
if you are using forever to keep your node app running, then typing forever list will show you the path to the log file that console.log is writing too
You can use the nodejs Console constructor
const mylog = new console.Console(
fs.createWriteStream("log/logger.log"),
fs.createWriteStream("log/error.log")
);
And then you can use it just like the normal console class, eg:
mylog.log("Ok!"); // Will be written into 'log/logger.log'
mylog.error("Bad!"); // Will be written into 'log/error.log'
I took on the idea of swapping the output stream to a my stream.
const LogLater = require ('./loglater.js');
var logfile=new LogLater( 'log'+( new Date().toISOString().replace(/[^a-zA-Z0-9]/g,'-') )+'.txt' );
var PassThrough = require('stream').PassThrough;
var myout= new PassThrough();
var wasout=console._stdout;
myout.on('data',(data)=>{logfile.dateline("\r\n"+data);wasout.write(data);});
console._stdout=myout;
var myerr= new PassThrough();
var waserr=console._stderr;
myerr.on('data',(data)=>{logfile.dateline("\r\n"+data);waserr.write(data);});
console._stderr=myerr;
loglater.js:
const fs = require('fs');
function LogLater(filename, noduplicates, interval) {
this.filename = filename || "loglater.txt";
this.arr = [];
this.timeout = false;
this.interval = interval || 1000;
this.noduplicates = noduplicates || true;
this.onsavetimeout_bind = this.onsavetimeout.bind(this);
this.lasttext = "";
process.on('exit',()=>{ if(this.timeout)clearTimeout(this.timeout);this.timeout=false; this.save(); })
}
LogLater.prototype = {
_log: function _log(text) {
this.arr.push(text);
if (!this.timeout) this.timeout = setTimeout(this.onsavetimeout_bind, this.interval);
},
text: function log(text, loglastline) {
if (this.noduplicates) {
if (this.lasttext === text) return;
this.lastline = text;
}
this._log(text);
},
line: function log(text, loglastline) {
if (this.noduplicates) {
if (this.lasttext === text) return;
this.lastline = text;
}
this._log(text + '\r\n');
},
dateline: function dateline(text) {
if (this.noduplicates) {
if (this.lasttext === text) return;
this.lastline = text;
}
this._log(((new Date()).toISOString()) + '\t' + text + '\r\n');
},
onsavetimeout: function onsavetimeout() {
this.timeout = false;
this.save();
},
save: function save() { fs.appendFile(this.filename, this.arr.splice(0, this.arr.length).join(''), function(err) { if (err) console.log(err.stack) }); }
}
module.exports = LogLater;
I just build a pack to do this, hope you like it ;)
https://www.npmjs.com/package/writelog
I for myself simply took the example from winston and added the log(...) method (because winston names it info(..):
Console.js:
"use strict"
// Include the logger module
const winston = require('winston');
const logger = winston.createLogger({
level: 'info',
format: winston.format.json(),
transports: [
//
// - Write to all logs with level `info` and below to `combined.log`
// - Write all logs error (and below) to `error.log`.
//
new winston.transports.File({ filename: 'error.log', level: 'error' }),
new winston.transports.File({ filename: 'combined.log' })
]
});
//
// If we're not in production then log to the `console` with the format:
// `${info.level}: ${info.message} JSON.stringify({ ...rest }) `
//
if (process.env.NODE_ENV !== 'production') {
logger.add(new winston.transports.Console({
format: winston.format.simple()
}));
}
// Add log command
logger.log=logger.info;
module.exports = logger;
Then simply use in your code:
const console = require('Console')
Now you can simply use the normal log functions in your file and it will create a file AND log it to your console (while debugging/developing). Because of if (process.env.NODE_ENV !== 'production') { (in case you want it also in production)...
Create a utils/logger.js file with:
var fs = require('fs');
var util = require('util');
var log_file = fs.createWriteStream(__dirname + '/../logs/server.log', { flags: 'w' });
var log_stdout = process.stdout;
console.log = function () { //
[...arguments].forEach(element => {
log_file.write(util.format(element) + '\n');
log_stdout.write(util.format(element) + '\n');
});
};
module.exports = {
console
}
Include the logger.js file from any file where you want to console.log like:
const console = require('./utils/logger').console;
Create a logs folder and create an empty server.log file in it and run your app :)
Rudy Huynh's solution worked really well for me. I added a little bit to have it spit out files with today's date and time.
var dateNow = new Date();
var timeNow = dateNow.getHours() + '-' + dateNow.getMinutes();
var logPath = "log/" + dateNow.toDateString() + ' -' + ' Start Time - ' + timeNow + ".log"
consoleLogToFile({
logFilePath: logPath
});
It's not very elegant but this way it'll save different, easy to read, log files instead of just updating the same "default.log" file.
Based on multiparameter version by Clément, just without color codes for the text file
var fs = require('fs');
var util = require('util');
var logFile = fs.createWriteStream('log.txt', { flags: 'a' });
// Or 'w' to truncate the file every time the process starts.
var logStdout = process.stdout;
console.log = function () {
// Storing without color codes
logFile.write(util.format.apply(null,arguments).replace(/\033\[[0-9;]*m/g,"") + '\n');
// Display normally, with colors to Stdout
logStdout.write(util.format.apply(null, arguments) + '\n');
}
Note: Answering since I couldn't comment