Current time: 10-18-2019, 02:09 PM Hello There, Guest! (LoginRegister)

Post Reply 
 
Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
VisuallyComplete, RUMFirstPaint, StartRender and SpeedIndex for mobile tests
10-20-2015, 06:23 AM
Post: #1
VisuallyComplete, RUMFirstPaint, StartRender and SpeedIndex for mobile tests
Hello everyone,

On my private instance, I get some values by calling jsonResult.php.
"Visually Complete"
Quote:jsonResult->data->median->firstView->visualComplete
jsonResult->data->median->repeatView->visualComplete
"RUM First Paint"
Quote:jsonResult->data->median->firstView->firstPaint
jsonResult->data->median->repeatView->firstPaint
"Start Render"
Quote:jsonResult->data->median->firstView->render
jsonResult->data->median->repeatView->render
"Speed Index"
Quote:jsonResult->data->median->firstView->SpeedIndex
jsonResult->data->median->repeatView->SpeedIndex
It seems webpagetest
1. always provides these 4 values as 0 for apple devices tests
2. sometimes provides some of these 4 values as 0 for android devices test

I'm not sure what's the reason....
If it's a bug, is it possible to provide both these 4 values for mobile devices? Or, what is the reason I got 0 as results?

On the attachment, I made screenshots of android tests results and iphone tests results. some columns(ex:RUM First Paint) are not shown when the value is 0.

Thanks,
   
Find all posts by this user
Quote this message in a reply
10-20-2015, 06:34 AM
Post: #2
RE: VisuallyComplete, RUMFirstPaint, StartRender and SpeedIndex for mobile tests
Do you have links to the actual test results? There are 2 iPhone agents. The old (soon to be decomissioned) Blaze agent running iOS 5 and the new agent that shares the same code as the android agent (the iPhone 5c running iOS 9).

iOS doesn't support the RUM first paint metric - though it's probably a bug that it reports 0 instead of -1

Speed Index, Visually Complete and Start render will only exist if video capture is enabled (and may not be reported with the old iOS agent). Also probably a bug if they report 0 instead of -1 when they are missing.
Visit this user's website Find all posts by this user
Quote this message in a reply
10-20-2015, 06:36 AM
Post: #3
RE: VisuallyComplete, RUMFirstPaint, StartRender and SpeedIndex for mobile tests
Sorry, I see it's a private instance - most of the comments still apply, just not the request for a link or the specifics about the iOS agent.

I'd highly recommend moving iOS to the new node agent in git. It's still a little rough and being worked on but it's orders of magnitude better than the blaze agent.
Visit this user's website Find all posts by this user
Quote this message in a reply
10-21-2015, 04:48 AM (This post was last modified: 10-21-2015 04:57 AM by matthew.)
Post: #4
RE: VisuallyComplete, RUMFirstPaint, StartRender and SpeedIndex for mobile tests
Thanks for your reminding on video capture, but the reason seems not that.
I check there is "video=1" in my url request
Quote:http://myserverurl/runtest.php?runs=5&video=1&url=http%3A%2F%2Fmytestedpage.com&location=Mobile_iPhone&f=json&k=mykey
I want to metion that I had put "--processvideo yes" in my command, hope it's not the reason.
Quote:/MyPathToScript/wptdriver.sh -m debug --browser iOS:40digitalserailnumber --serverUrl http://myserverUrl.net --location Mobile_iPhone --processvideo yes --apiKey myapikey
(path, serial number, server url and apikey are just fakeTongue)



I think the agent I'm running is the new agent code that you recommended:
Quote:git remote -v
origin http://github.com/WPO-Foundation/webpagetest.git (fetch)
origin http://github.com/WPO-Foundation/webpagetest.git (push)

git status
On branch master
Your branch is up-to-date with 'origin/master'.

Is my setting correct?

And another issue I also want to ask (I guess it has some relations with the video capture):
sometimes, the screen shots of ios devices can't shown up, I checked on the host, and there is no screen shot pictures there. Any ideas why?
thanks,
   

"--processvideo yes" is not the reason, it doesn't help even I remove this parameter
Find all posts by this user
Quote this message in a reply
10-22-2015, 12:42 AM
Post: #5
RE: VisuallyComplete, RUMFirstPaint, StartRender and SpeedIndex for mobile tests
It turns out git pull again can solve everything Smile
Find all posts by this user
Quote this message in a reply
10-22-2015, 04:06 AM
Post: #6
RE: VisuallyComplete, RUMFirstPaint, StartRender and SpeedIndex for mobile tests
To say iOS support is in a state of flux would be an understatement (particularly around video capture). I've been working on it heavily for the past few weeks and most of the kinks are worked out but there are still a lot of painful issues (the most painful of which is that OSX will only record video from one device at a time which makes scaling really expensive).

I'm planning on poking at using VNC to do video capture as an option which would remove the whole OSX requirement and allow for a lot more devices per host. Right now I have it set up so that 2 devices can share a single host reasonably well with them blocking each other only at the actual testing part where video is recorded (video processing, device prep, etc can happen while the other device is testing).
Visit this user's website Find all posts by this user
Quote this message in a reply
10-22-2015, 09:49 AM
Post: #7
RE: VisuallyComplete, RUMFirstPaint, StartRender and SpeedIndex for mobile tests
(10-22-2015 04:06 AM)pmeenan Wrote:  To say iOS support is in a state of flux would be an understatement (particularly around video capture). I've been working on it heavily for the past few weeks and most of the kinks are worked out but there are still a lot of painful issues (the most painful of which is that OSX will only record video from one device at a time which makes scaling really expensive).

I'm planning on poking at using VNC to do video capture as an option which would remove the whole OSX requirement and allow for a lot more devices per host. Right now I have it set up so that 2 devices can share a single host reasonably well with them blocking each other only at the actual testing part where video is recorded (video processing, device prep, etc can happen while the other device is testing).

It sounds really a huge improvement, thanks for your hard working!
Find all posts by this user
Quote this message in a reply
Post Reply 


Forum Jump:


User(s) browsing this thread: 2 Guest(s)