How can I combine the output and error logs?

[dandv]

forever has a -l option that directs both STDOUT and STDERR of the app into the same file. This is useful as I can see after which operations (output via console.log) some error has occurred (output via console.error).

How can I do this with pm2?

[soyuka]

I don't think it's a good idea but perhaps that setting the same output file into a JSON declaration would work ;).

[dandv]

I tried passing the same output file to -o and -e, but pm2 logs was confused by that. Every STDOUT line was shown both as "out" and as "err".

[Unitech]

You can now merge all logs of the same process name to the same output and error file via the option --merge-logs

[dandv]

Let's say I only have one process. Can --merge-logs interleave the -e (error) and -o (output) logs into one file? This would assist in seeing what has happened before an error occurred, because the error message itself in a separate file will be out of context.

[dandv]

Could we maybe reopen this issue? As I see, --merge-logs combines logs from instanced, but not STDOUT and STDERR.

Here's a workaround I found, until pm2 enables support for this. It's a bit more complicated because of the numbers added to the end of the log filenames.

  MYLOG_BASENAME="myapp.log"
  pm2 start myapp.js -o $MYLOG_BASENAME.out -e $MYLOG_BASENAME.err
  ACTUAL_LOG=$(pm2 -m l | grep -o /.\*$MYLOG_BASENAME.\*out)
  ACTUAL_ERR=$(pm2 -m l | grep -o /.\*$MYLOG_BASENAME.\*err)
  tail -f $ACTUAL_LOG $ACTUAL_ERR | tee -a $MYLOG_BASENAME.both

Please let me know if there's a simpler way to achieve this.

[dandv]

Any updates? The workaround above require running tail all the time, in parallel with the app. Cumbersome.

[liukun]

Seems no good options here. My best bet is to use Heka or Logstash listening on TCP to manage logs yourself.

[soyuka]

#1996