Browser Metrics:Data Collectors

Data Format

Overview

Data will be collected as a stream of events, with the exception of a user profile that is collected once every browser session. Each event type is represented by a particular XML element in the http://www.mozilla.org/metrics namespace. We will define several event types that will have predefined schemas. We will be able to throttle data collection based on the event type, so that verbose events can be turned on and off when we want. The tentative event types and associated schemas are listed below. (Note: the list is still very much in flux and will likely change.)

All event types have a time attribute, which gives a timestamp for the event, recorded as seconds since the epoch.

Profile

The profile will be collected every browser session, to pick up changes periodically. The profile element is a grouping element which contains a set of profile values.

Element <profile>

Attributes:

session = [integer]

Gives the current session count, incremented everytime the browser loads

time = [integer]

Gives the unixtime when the profile was captured

The following elements are optional children of <profile>:

Element <bookmarks>

A container element for 1 or more bookmarklocation elements.

Attributes: none

Element <bookmarklocation>

Contains information about the location of bookmarks

Attributes:

foldercount = [integer]

Gives the number of bookmark folders

itemcount = [integer]

Gives the total number of bookmark items

livemarkcount = [integer]

Gives the total number of livemarks

separatorcount = [integer]

Gives the total number of separators

name = [string]

Gives the name of the set of bookmarks (full-tree, root, toolbar)

Element <cpu>

Describes the CPU of the client system.

Attributes:

arch = [string]

Gives the CPU architecture (x86, powerpc, etc).

Element <display>

Describes the user's display.

Attributes:

xsize = [integer]

Gives the number of pixels horizontally on the primary screen.

ysize = [integer]

Gives the number of pixels vertically on the primary screen.

screens = [integer]

Gives the number of screens on the system.

Element <extensions>

A container element for 1 or more extension elements.

Attributes: none

Element <extension>

Contains information about a single extension.

Attributes:

extensionid = [string]

Gives an MD5 hash of the extension's id.

version = [string]

Gives the extension's version number.

Element <install>

Describes the installation of Firefox the user is running.

Attributes:

appversion = [string]

Gives the current application version (e.g. 3.0b4)

buildid = [integer]

Gives the Build ID (e.g. 20060327).

default = [boolean]

Whether the browser is set as the default browser for the current user.

extversion = [string]

Gives the current spectator extension version (e.g. 2.1a9)

installdate = [integer]

Gives the date the build was first run on this profile (seconds since the epoch).

locale = [string]

Gives the locale (e.g. en-US, de, etc)

Element <memory>

Describes the memory configuration of the client system.

Attributes:

mb = [integer]

Gives the number of megabytes of system memory.

Element <os>

Describes the operating system the client is running.

Attributes:

name = [string]

Gives the name of the operating system, e.g. "Windows" or "MacOSX".

version = [string]

Gives the version of the operating system, e.g. "XP" or "Tiger".

Element <plugins>

A container element for 1 or more plugin elements.

Attributes: none

Element <plugin>

Contains information about a single plugin.

name = [string]

Gives an MD5 hash of the plugin name.

version = [string]

Gives the version of the plugin.

Load Event

Load events record a document being loaded into a DOM Window.

Element <document>

Attributes:

action = [string]

Records if the document is being loaded or destroyed

window = [integer]

The id of the window where the document was loaded.

bfCacheHit = [boolean] (optional)

Whether the document presentation was loaded from the session history cache. If not specified, assumed to be false.

docid = [integer]

The id of the document that was loaded

loadtime = [integer]

The time from the initiation of the load until the document is complete (which includes all images, stylesheet, etc) in milliseconds.

memresident = [integer]

The resident memory after the page has loaded

memtotal = [integer]

The total memory usage after the page has loaded

origin = [string] (optional)

The action which initiated the load. Possible values include:

• typed: The document URI was typed (or pasted) by the user.
• link: The user followed a link to the document URI.
• session-history: The user used back/forward navigation to load the document.
• reload: The user used the reload button or keyboard shortcut to reload the document.
• global-history (not implemented): The user loaded the page by selecting it from their global history.
• bookmark (not implemented): The user loaded the page by selecting it from the bookmarks menu, bookmarks toolbar, or bookmarks management UI.
• script (not implemented): A script executing on a page loaded the document.
• refresh: A meta-refresh loaded the document.
• external (not implemented): The document URI was passed in from an external application.

session = [integer]

Gives the current session count, incremented everytime the browser loads

subframe = [integer]

Whether the document was loaded from a subframe. If not specified, assumed to be false.

time = [integer]

Gives the unixtime when the document load was captured

urlhash = [string]

Gives the md5 hash of chrome document load URLs

UI Event

Records UI events

Element <uielement>

Attributes:

action = [string]

Records what caused the event (command, popupshowing)

keyidhash = [string]

Records the md5 hash of the key combo that caused the event (goBackKb, etc)

session = [integer]

Gives the current session count, incremented everytime the browser loads

targetanonidhash = [string]

Gives the md5 hash of the anonymous target id that caused the event

targetidhash = [string]

Gives the md5 hash of the target id that caused the event

time = [integer]

Gives the unixtime when the event was captured

window = [integer]

The id of the window where the document was loaded.

Other Ideas

Garbage Collection

• number/size of objects reachable
• time spent doing garbage collection

Startup/Shutdown

• time taken

Places UI Event

We'll have to be sure not to duplicate data here between the UI Event and the Places UI Event. This could potentially be an extension of the regular UI Event.

• node types expanded/clicked
• periods of time searched

Autocomplete

We may try to fit this into the schema of another event type.

• how far down the list?
• match but typed anyway?

Open Issues

• We may want to consolidate some of these events into summary statistics, to compact the amount of data we are collecting. However, this may not be necessary since we will be able to throttle the data collection on a per event type.

• The event types listed above are still in flux and may be combined or removed.