1525494 - load all DiscoveryStream component feeds in parallel rather than serially

Status: VERIFIED FIXED

(Firefox :: New Tab Page, defect, P1)

Description

[Dan Mosedale (:dmosedale, :dmose)]

Instead of loading all the feeds one after another, we should load them in parallel so that they finish sooner and things display sooner.

Comment 1

[GitHub Bugzilla PR Linker]

Attachment: Link to GitHub pull-request: https://github.com/mozilla/activity-stream/pull/4768

Comment 2

[Scott [:thecount] Downe]

[Tracking Requested - why for this release]: Performance fixes so we don't regress what it's like on release for the test.

Comment 3

[Kate Hudson :k88hudson]

Attachment: GitHub Pull Request

Comment 4

[Kate Hudson :k88hudson]

https://github.com/mozilla/activity-stream/commit/84ef9fa4825ae3368f4a76d711468c6d96c74555

Comment 5

[Ed Lee :Mardak]

https://hg.mozilla.org/mozilla-central/rev/6d9b468951ce

Comment 6

[Kate Hudson :k88hudson]

Attachment: Bug 1525494 - Parallel component feed loads

Comment 7

[Kate Hudson :k88hudson]

Can you add the uplift request / test steps for this and ask QA to verify?

Comment 8

[Scott [:thecount] Downe]

Comment on attachment 9043736 [details]
Bug 1525494 - Parallel component feed loads

Beta/Release Uplift Approval Request

Feature/Bug causing the regression

None

User impact if declined

Experiment results and user experience because of performance.

Is this code covered by automated tests?

Yes

Has the fix been verified in Nightly?

No

Needs manual test from QE?

Yes

If yes, steps to reproduce

1. Set the pref browser.newtabpage.activity-stream.discoverystream.config to {"enabled":true,"show_spocs":false,"layout_endpoint":"https://getpocket.com/v3/newtab/layout?version=1&consumer_key=40249-e88c401e1b1f2242d9e441c4&layout_variant=dev-test-feeds"}
2. Load a new tab.

It should load a few seconds faster than before the patch. In my tests I got around 18 seconds without the patch and 10 with, which is pretty noticeable. That link is designed to be slow as it loads a bunch of feeds. There is caching so the second attempt should be significantly faster.

List of other uplifts needed

None

Risk to taking this patch

Medium

Why is the change risky/not risky? (and alternatives if risky)

Behind pref

String changes made/needed

None

Comment 9

[Brahmini Nagabandi]

QA Results:

Tested on :

FF Nightly version : 67.0a1 (2019-02-15) - This is the latest build while I was testing.
OS : Mac and Windows 10 Pro

Mac OS:
I did not notice much delay for the feeds to load. Response time was ~3seconds, and 2nd attempt was pretty fast.

Windows OS:
I could see the new tab loaded after 12seconds. 2nd attempt was immediate.

Comment 10

[Scott [:thecount] Downe]

I think this is expected results. 12 seconds seems roughly close to the expected time of 10 seconds.

Comment 11

[Brahmini Nagabandi]

Thanks Scott.

Additional note. I have this Windows 10 Pro on my VM.

Works as expected.

Marking as verified.

Comment 12

[Liz Henry (:lizzard) (relman/hg->git project)]

Comment on attachment 9043736 [details]
Bug 1525494 - Parallel component feed loads

Planned work for pocket/new tab. Verified in Nightly.
OK for beta uplift, should land for beta 9.
Landing order: bug 1519879, bug 1525494, bug 1526861, bug 1524669, bug 1527195, bug 1525391, bug 1527347, bug 1525366, bug 1527626, bug 1527397, bug 1518258, bug 1527701, bug 1527370.

Comment 13

[Cosmin Sabou [:CosminS]]

https://hg.mozilla.org/releases/mozilla-beta/rev/3c371055b381

Comment 14

[Camelia Badau [:cbadau], Desktop Test Engineering]

(In reply to Brahmini Nagabandi from comment #11)

Thanks Scott.

Additional note. I have this Windows 10 Pro on my VM.

Works as expected.

Marking as verified.

Can you please verify this issue on Firefox 66 Beta 9 (https://archive.mozilla.org/pub/firefox/candidates/66.0b9-candidates/build1/)?

Comment 15

[Brahmini Nagabandi]

BETA Testing:

QA Results:

Tested on :

FF Beta version : 66.0b9 (64-bit)
OS : Mac and Windows 10 Pro
Date : Feb 21

**Mac OS:
I did notice the response time was ~6seconds, and 2nd attempt was pretty fast.

Windows OS:
I could see the new tab loaded at ~20seconds. 2nd attempt was fast and took ~5seconds.

Scott - can you confirm if this is expected?

Comment 16

[Scott [:thecount] Downe]

Yeah that sounds reasonable for that layout/feed.

Comment 17

[Camelia Badau [:cbadau], Desktop Test Engineering]

Marking the status-firefox66 tracking flag as verified based on previous comments.

Thanks for verifying this as well!

Comment 18

[Brahmini Nagabandi]

(In reply to Camelia Badau [:cbadau], Release Desktop QA from comment #17)

Marking the status-firefox66 tracking flag as verified based on previous comments.

Thanks for verifying this as well!

Thank you.