blob: 937925fba0ee9221e39f9cebfdd4eca4ce8aa840 [file] [log] [blame] [view]
Alex Mineer2be39f72017-06-13 16:21:301# Release Blockers
2
3[TOC]
4
5## tl;dr
6
7* Only mark bugs as blockers if the product **must not** be shipped with the
8 bug present.
9* **Everyone** on the team can add or remove blocking labels.
10* Evaluate bugs as potential blockers based on their **severity** and
11 **prevalence**.
12* Provide **detailed rationale** whenever adding or removing a blocking label.
13* Ensure all blockers have an **OS and milestone** tagged.
14* Release owners have final say on blocking status; contact them with any
15 questions.
16
17## Context
18
19The Chromium project utilizes release block labels to help define each
20milestone's critical path. The following labels are available:
21
22* **ReleaseBlock-Dev**, which blocks shipping to the dev, beta, and stable
23 channels.
24* **ReleaseBlock-Beta**, which blocks shipping to the beta and stable
25 channels.
26* **ReleaseBlock-Stable**, which blocks shipping to the stable channel.
27
28Release block labels must be used in conjunction with a milestone (M=##) label
29as well as an OS (e.g. OS-Android). The combination of blocking label,
30milestone, and OS determine which releases we hold; e.g. OS=Android M=59
31ReleaseBlock-Beta means we will not ship M59 Android builds to either the beta
32or stable channel until the bug is addressed. Android M59 dev releases, or
33releases on any other platform, would not be blocked.
34
35Because these labels are used to manage critical path, they should **not** be
36used unless we genuinely believe we cannot ship to a given channel with the
37issue present. Do not mark an issue as a blocker "to ensure someone looks at
38it." Priority and milestone labels should be used for this purpose.
39
40These rules apply to bugs and regressions; teams may use criteria for blocking
41due to other work (e.g. tasks) as they see fit.
42
43## Assessing Blockers
44
45Issues should be evaluated for release blocking status using the following
46matrix based on the issue's severity and prevalence:
47
48| | Low Impact | Medium | High | Critical |
49|---------------|---------------------|---------------------|---------------------|------------------|
50| **Few Users** | | | ReleaseBlock-Stable | ReleaseBlock-Dev |
51| **Some** | | ReleaseBlock-Stable | ReleaseBlock-Beta | ReleaseBlock-Dev |
52| **Most** | ReleaseBlock-Stable | ReleaseBlock-Beta | ReleaseBlock-Beta | ReleaseBlock-Dev |
53| **All** | ReleaseBlock-Stable | ReleaseBlock-Beta | ReleaseBlock-Dev | ReleaseBlock-Dev |
54
55### Severity
56
57Severity is defined as the impact to a user who experiences the bug.
58
59* **Critical**: A bug with extreme consequence to the user, e.g. a regression
60 in privacy (leaking user data), loss of user data, crash on startup, etc.
61 These bugs must be fixed immediately and thus should block any release where
62 they are present.
63* **High**: A bug with large impact to the user, e.g. a CSS rendering issue
64 causing content to disappear, videos not playing, extreme jank, etc. There
65 is no simple workaround for the issue.
66* **Medium**: A bug with moderate impact to the user, e.g. a CSS rendering
67 issue causing content to be misaligned, moderate jank, non-startup crash,
68 memory regressions, etc. There may be a workaround for the issue.
69* **Low**: A bug with little impact to the user, generally cosmetic in nature
70 and easy to work around.
71
72### Prevalence
73
74Prevalence is defined as the volume of users who will experience the bug.
75* **Few Users (<5%)**: The bug requires many steps to trigger, or is dependent
76 on timing, e.g. two simultaneous taps on different parts of the screen.
77* **Some (5% - 35%)**: The bug affects a minor workflow, or requires a series
78 of steps to trigger.
79* **Most (35% - 75%)**: The bug affects a major workflow, e.g. sync,
80 downloading files, etc.
81* **All (75% - 100%)**: The bug affects core product functionality, e.g.
82 scrolling a page.
83
84Note that prevalence should be evaluated based on the population of users they
85affect - e.g. a bug affecting all Android users (but not Windows users)
86would still be considered to affect all users, and a bug affecting all
87Enterprise Windows users (but not all consumer Windows users) could also be
88considered to affect all users.
89
90### Customization
91
92The definitions provided above are examples; teams are encouraged to customize
93where it makes sense, e.g. the web platform team may consider developer impact
94for severity and feature usage for prevalence.
95
96## Blocker Management
97
98**Everyone should feel free to add, modify, or remove release blocking labels
99where appropriate, so long as you follow the guidelines below.** If a TPM or
100test engineer has marked a bug as a release blocker, but a developer knows for
101sure that the issue should not block the release, the developer should remove
102the release blocking label; similarly anyone should feel free to add a release
103blocking label to a bug they feel warrants holding a release. That said, there
104are some general guidelines to follow:
105
106* Be specific and descriptive in your comments when tagging, or untagging, an
107 issue as a release blocker. **You must explain your reason for doing so.**
108 Including your rationale around impact and prevalence will make it much
109 easier for anyone reviewing the bug to understand why the bug should (or
110 should not) block the release. It will also help anyone re-assessing the
111 bug if we receive new information later.
112* **When in doubt, be conservative and mark bugs as blockers!** It's better
113 to tag a bug as a release blocking issue and have the label removed later
114 than to ship the bug to users and have to respin due to unanticipated
115 consequences. You can always loop in the release management team (by CC'ing
116 onto the bug, pinging, or e-mailing) the platform owners listed on the
117 [Chrome Calendar](https://chromepmo.appspot.com/calendar) (Googlers only,
118 opening to all in the near future) for their input if you need assistance.
119* These guidelines should be used with the data we have available at the time.
120 **If we need more data to make a good decision, get help in finding it!**
121 When new information arises, re-assess the bug using the new details as soon
122 as possible.
123* The release management team (release engineers and TPMs) have the final say
124 when it comes to release blocking issues - they can tag, and untag, issues
125 as they see fit.
126
127## Other Considerations
128
129### New Features
130
131Any bugs related to a feature that is behind a flag and is not enabled by
132default should never block a release (they should be disabled instead).
133Features enabled by default should follow the proposals listed above.
134
135### Regressions
136
137Regressions should follow the same guidelines as listed above; an issue should
138not be tagged as a release blocker simply because it is a regression. While
139we'd like to prevent regressions in general, there is a large backlog of bugs we
140need to address, and we should focus on the most important. To ensure we
141maintain a high bar for product quality, we should track the number of
142introduced versus escaped regressions, and follow up if the number starts to
143rise.