Console Log UI -- One Request Broken Up Into Several Expandable Rows

[Andrei Matveev]

Hello,

Have there been any recent changes to developer console's Logs UI? 

Within the past day or two I've started seeing logs of long running requests (i.e. ones w/ lots of log entries) get split up into several separate expandable columns - i.e. the full log of a request can no longer be seen by expanding  a single row (instead I have to expand several consecutive logs to see the full request log).

I suspect that it's something to do w/ the Logs UI (and not a code change on my side) b/c when I go back to look at logs from 3-4 weeks ago they also appear fragmented. I know for a fact that I've previously been able to look at those logs in non-fragmented form.

fyi I'm still new to App Engine so it's totally possible that I unwittingly toggled some viewing option that's causing this... in that case I apologize in advance for the dumb question :)

[Patrice (Cloud Platform Support)]

Hi Andrei,

Do you have a print screen of this? I can't seem to see the same in my own console, no matter how many logs I push into one requests.

Thanks

[pdknsk]

I noticed this too. 

[Andrei Matveev]

Apologies for the reply delay...

Below are screenshots that should help paint a better picture of what I'm talking about. Please excuse all the blurring... still operating in stealth mode.

Fragmented request in the console log UI

Beginning of request:

Continuation of request:

Same Log Fragment in the Old App Engine Log UI:

On Friday, September 25, 2015 at 4:43:43 PM UTC-4, pdknsk wrote:

I noticed this too. 

[Patrice (Cloud Platform Support)]

Hey Andrei,

Thank you very much for all this. I've made some more research, and we've seen it happen, but for extremely numerous logs (>100K). It is the first time I hear of this happening with a request that only contains 10 entries.

And no worry for the blurr, I just wanted to see the look of your request. If we need an un-blurred print screen, I'll let you know, and you can always just reply to me privately with it to keep it off-thread. For now it won't be needed.

I'll keep on checking on my end and come back with what I find.

Cheers!

[pdknsk]

In the first screenshot, a status note is logged every second. I guess the second screenshot speaks for itself. This is very rare.

I've noticed some other changes. I think it used to be that a request was logged when it finished. Now a request is logged in the slot it started in. This can cause requests to be swallowed when using continues logging. Let me explain.

This is with continues logging enabled.

00:50:00 10ms /

00:30:00 10ms /

00:10:00 10ms /

00:00:00 10ms /

This is when you refresh the log page.

00:50:00 10ms /

00:30:00 10ms /

00:20:00 20s /www/

00:10:00 10ms /

00:00:00 10ms /

Suddenly another request appears. So what happened here. At 00:20:00, request /www/ started processing, but took 20s and thus only finished at 00:40:00, at which point another request was already logged. This causes request /www/ not to be logged in continuous logging.

I think this is how it used to be.

00:50:00 10ms /

00:40:00 20s /www/

00:30:00 10ms /

00:10:00 10ms /

00:00:00 10ms /

Request /www/ is logged when it finishes, and not in the slot it started.

[pdknsk]

PS. By continuous logging I mean streaming logs.

[Andrei Matveev]

Ping. Any updates on this?

[Patrice (Cloud Platform Support)]

Hi Andrei,

Sorry for not keeping you posted here. We're actually still investigating after a couple of possible avenues for this, and as soon as I find something, I will let you know.

Thank you very much for your patience in the meantime.

Cheers!

[PK]

I run into this today as well in the new UI. Very annoying, I hope it gets fixed soon. 

--PK

--
You received this message because you are subscribed to the Google Groups "Google App Engine" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-appengi...@googlegroups.com.
To post to this group, send email to google-a...@googlegroups.com.
Visit this group at http://groups.google.com/group/google-appengine.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-appengine/05dd5024-9749-4c35-aefa-0b440b75588f%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Andrei Matveev]

Ping

Is there any update on this? It's getting exceedingly hard to work around the issue... especially w/ the steady deprecation of parity features in the old App Engine console (e.g. task queues, cron-job, etc. already redirect to the google cloud dev console).

[Patrice (Cloud Platform Support)]

Hi all,

So, even if we unfortunately have no date for a possible fix yet, it does seem that what's happening is explainable :

As you guys may assume, the logs are collected into buffers and then the buffers concatenated together to be displayed. It seems that, even though the "collected into buffers" happens the same way in the new console as it did in the old, the concatenation seems to be not behaving consistently, causing for these issues to spring up.

Have a great weekend!

[Patrice (Cloud Platform Support)]

Hi again all,

First, I'd like to apologize if I was unclear in my last message. Even if this is explainable, I didn't mean to imply this isn't an issue. I simply meant that we are progressing on finding a root cause as to why this is happening. Rereading my previous message, I can definitely understand how it seemed to imply "all is good" when in fact it most definitely isn't.

This thread did indicate there is a clear issue with the logs viewer, and we are actively working on fixing it. To properly track this, I went ahead and filed this issue, which the relevant back-end team is aware of. I would suggest starring this issue so you can be made aware of the progress by the back-end team.

Cheers!