_posts/2019-07-26-gerrit-news-jun-jul-2019.md - homepage - Git at Google

 ---
 title: "Gerrit Project News #3: June-July 2019"
 tags: news
 keywords: news
 permalink: 2019-07-26-gerrit-news-jun-jul-2019.html
 summary: "Gerrit project news from June and July 2019."
 hide_sidebar: true
 hide_navtoggle: true
 toc: true
 ---

 ## Community Managers

 The Gerrit Community Managers for 2019 were
 [announced](https://www.gerritcodereview.com/2019-06-28-cm-announce.html):

 * Edwin Kempin (Google)
 * Marco Miller (Ericsson)
 * Matthias Sohn (SAP)

 See the linked post for details.

 ## User Summits and Hackathons in Sweden and USA

 The schedules are finalized for the user summits in
 [Gothenburg](https://gerrit.googlesource.com/summit/2019/+/refs/heads/master/schedule-europe.md)
 and [Sunnyvale](https://gerrit.googlesource.com/summit/2019/+/refs/heads/master/schedule-usa.md).

 ## Work on Gerrit Performance

 In 2019, one of the major efforts we are working on is performance. Gerrit’s new
 UI has support for client-side monitoring through plugins which is great for
 measuring exactly what the user experiences. We use these client side metrics
 and apply the
 [RAIL](https://developers.google.com/web/fundamentals/performance/rail) model to
 provide targets for latency. The current focus is on read latency - that is
 browsing around on the Gerrit web UI and looking at changes, diffs and
 dashboards. Write latency - that is posting a code review, creating draft
 comments, ... - will follow once reads are acceptably fast.

 This is an effort that goes all across the stack, it covers the UI, Gerrit’s
 backend as well as storage formats of data on disk where necessary.

 Until now we picked a lot of low hanging fruits that speed things up notably.
 For example: Rendering Polymer components
 lazyly instead of all-at-once when the app starts up; sending index queries in
 parallel if the index backend supports that; inline data that is required when
 the app starts into index.html to spare extra requests; rendering the correct
 avatar sizes by default to spare extra round trips to the backend.

 In the upcoming months, we want to fine-tune caching, work on external IDs and
 have the UI send requests to the backend earlier to allow for longer processing
 without latency impact.
	---
	title: "Gerrit Project News #3: June-July 2019"
	tags: news
	keywords: news
	permalink: 2019-07-26-gerrit-news-jun-jul-2019.html
	summary: "Gerrit project news from June and July 2019."
	hide_sidebar: true
	hide_navtoggle: true
	toc: true
	---

	## Community Managers

	The Gerrit Community Managers for 2019 were
	[announced](https://www.gerritcodereview.com/2019-06-28-cm-announce.html):

	* Edwin Kempin (Google)
	* Marco Miller (Ericsson)
	* Matthias Sohn (SAP)

	See the linked post for details.

	## User Summits and Hackathons in Sweden and USA

	The schedules are finalized for the user summits in
	[Gothenburg](https://gerrit.googlesource.com/summit/2019/+/refs/heads/master/schedule-europe.md)
	and [Sunnyvale](https://gerrit.googlesource.com/summit/2019/+/refs/heads/master/schedule-usa.md).

	## Work on Gerrit Performance

	In 2019, one of the major efforts we are working on is performance. Gerrit’s new
	UI has support for client-side monitoring through plugins which is great for
	measuring exactly what the user experiences. We use these client side metrics
	and apply the
	[RAIL](https://developers.google.com/web/fundamentals/performance/rail) model to
	provide targets for latency. The current focus is on read latency - that is
	browsing around on the Gerrit web UI and looking at changes, diffs and
	dashboards. Write latency - that is posting a code review, creating draft
	comments, ... - will follow once reads are acceptably fast.

	This is an effort that goes all across the stack, it covers the UI, Gerrit’s
	backend as well as storage formats of data on disk where necessary.

	Until now we picked a lot of low hanging fruits that speed things up notably.
	For example: Rendering Polymer components
	lazyly instead of all-at-once when the app starts up; sending index queries in
	parallel if the index backend supports that; inline data that is required when
	the app starts into index.html to spare extra requests; rendering the correct
	avatar sizes by default to spare extra round trips to the backend.

	In the upcoming months, we want to fine-tune caching, work on external IDs and
	have the UI send requests to the backend earlier to allow for longer processing
	without latency impact.