|
| 1 | +# Release Rollout Monitoring |
| 2 | + |
| 3 | +Use this checklist after a Play Console submission is ready for review or a |
| 4 | +release has been approved for tester or production rollout. It assigns release |
| 5 | +ownership, defines rollout gates, and lists the signals the team must watch |
| 6 | +before widening availability. |
| 7 | + |
| 8 | +## Ownership |
| 9 | + |
| 10 | +- Release owner: assign one engineer before submission. This person owns Play |
| 11 | + Console status checks, rollout decisions, incident coordination, and final |
| 12 | + go/no-go notes. |
| 13 | +- Backup owner: assign one alternate before submission. This person must have |
| 14 | + access to the same dashboards and can pause rollout or coordinate rollback if |
| 15 | + the release owner is unavailable. |
| 16 | +- Handoff channel: record the owner, backup, release version, build number, |
| 17 | + release track, and expected monitoring windows in the team release channel or |
| 18 | + shared release note. |
| 19 | +- Required access: both owners need Google Play Console access, Firebase console |
| 20 | + access for the `ontime-c63f1` project, GitHub Actions access, and access to |
| 21 | + production backend logs or dashboards. |
| 22 | + |
| 23 | +## Rollout Stages |
| 24 | + |
| 25 | +1. Internal testing |
| 26 | + - Upload the signed AAB to Play Internal Testing as a draft release. |
| 27 | + - Confirm install, login, notification permission flow, Firebase |
| 28 | + initialization, FCM token registration, and the highest-risk user flows on |
| 29 | + at least one real Android device. |
| 30 | + - Do not proceed if install fails, Firebase initialization fails, login is |
| 31 | + broken, or the backend receives no token registration for a fresh install. |
| 32 | +2. Closed testing or limited tester rollout |
| 33 | + - Release to the agreed tester group after internal testing is clear. |
| 34 | + - Hold for at least one active test pass or one business day, whichever is |
| 35 | + more useful for the release risk. |
| 36 | + - Do not proceed if testers report launch blockers, sign-in blockers, alarm |
| 37 | + delivery regressions, or data-loss behavior. |
| 38 | +3. Production staged rollout |
| 39 | + - Start with the smallest practical Play production percentage. |
| 40 | + - Increase only after the first 24-hour checks are clean and no policy, |
| 41 | + crash, ANR, backend, or review signal suggests a release-caused regression. |
| 42 | + - Record every percentage increase with timestamp, owner, build number, and |
| 43 | + the dashboards checked. |
| 44 | +4. Full rollout |
| 45 | + - Move to full rollout only after the first 7-day checks are clean or the |
| 46 | + release owner explicitly accepts the remaining known risk. |
| 47 | + - Keep monitoring for one additional business day after reaching 100%. |
| 48 | + |
| 49 | +## Pause And Rollback Criteria |
| 50 | + |
| 51 | +Pause the rollout immediately when any of these signals appear release-caused: |
| 52 | + |
| 53 | +- Play Console review rejection, policy warning, app content warning, or account |
| 54 | + email requiring action. |
| 55 | +- Crash rate or ANR rate materially above the previous production baseline. |
| 56 | +- Repeated startup, login, schedule creation, notification, alarm, or token |
| 57 | + registration failures from testers or monitoring. |
| 58 | +- Backend error spikes for authentication, schedule, preparation, alarm, or FCM |
| 59 | + token endpoints after the new build reaches users. |
| 60 | +- Ratings or reviews identify a reproducible blocker in the new version. |
| 61 | +- Support or team reports indicate data loss, missed alarms, broken sign-in, or |
| 62 | + privacy-sensitive behavior. |
| 63 | + |
| 64 | +Rollback or supersede the release when pausing is not enough: |
| 65 | + |
| 66 | +- If the issue is server-side and a backend fix can safely restore behavior, |
| 67 | + deploy the backend fix and keep the mobile rollout paused until metrics |
| 68 | + recover. |
| 69 | +- If the issue is client-side and affects already-updated users, prepare a new |
| 70 | + fixed build and keep rollout paused until it is approved. |
| 71 | +- If Play Console allows halting the staged rollout before wide exposure, halt |
| 72 | + it and document the affected percentage and build number. |
| 73 | +- If a policy issue caused rejection or removal risk, follow the rejection |
| 74 | + response playbook from #461 once available and do not resubmit until the facts, |
| 75 | + fix, and declarations are aligned. |
| 76 | + |
| 77 | +## Monitoring Locations |
| 78 | + |
| 79 | +- Play Console review status: check the release status, publishing overview, app |
| 80 | + content warnings, policy status, release notes, and tester or production track |
| 81 | + rollout percentage. |
| 82 | +- Play Console Android Vitals: monitor crash rate, user-perceived crash rate, |
| 83 | + ANR rate, excessive wakeups, battery, and device-specific clusters. |
| 84 | +- Play Store ratings and reviews: watch new public reviews, tester feedback, and |
| 85 | + low-rating comments that mention the current version or rollout date. |
| 86 | +- Policy email: monitor the Google Play developer account inbox and any team |
| 87 | + forwarding address for rejection, declaration, data safety, or policy notices. |
| 88 | +- Firebase console: check Cloud Messaging delivery health and any crash or event |
| 89 | + dashboard the project has enabled for `ontime-c63f1`. |
| 90 | +- GitHub Actions: confirm the release workflow artifact, generated build number, |
| 91 | + and deploy job match the build that reached Play Console. |
| 92 | +- Backend monitoring: check production API logs and dashboards for elevated |
| 93 | + status codes, latency, authentication failures, schedule/preparation failures, |
| 94 | + alarm status report failures, and FCM token registration failures. |
| 95 | +- Team channels and support inbox: watch for tester reports, screenshots, device |
| 96 | + details, and reproduction steps. |
| 97 | + |
| 98 | +## First 24 Hours |
| 99 | + |
| 100 | +- At submission: record version name, generated build number, track, rollout |
| 101 | + percentage, release owner, backup owner, and expected next check time. |
| 102 | +- Every 2 to 4 hours during local business hours: check Play review status, |
| 103 | + policy email, Android Vitals, Firebase health, backend errors, and team |
| 104 | + reports. |
| 105 | +- After approval or rollout start: verify installs from the active track and run |
| 106 | + a smoke test on a real Android device. |
| 107 | +- Before any percentage increase: compare crashes, ANRs, backend errors, and |
| 108 | + tester or review feedback against the pre-rollout baseline. |
| 109 | +- End of day: post a short status note with current rollout percentage, checked |
| 110 | + dashboards, issues found, and the next planned decision. |
| 111 | + |
| 112 | +## First 7 Days |
| 113 | + |
| 114 | +- Check Play Console, policy email, Android Vitals, reviews, Firebase, and |
| 115 | + backend dashboards at least once per business day. |
| 116 | +- Track whether crash and ANR clusters are new, increasing, or tied to a |
| 117 | + specific device, OS version, or app version. |
| 118 | +- Review low-rating feedback and tester reports for repeated symptoms before |
| 119 | + widening rollout. |
| 120 | +- Confirm backend errors and latency remain stable for schedule, preparation, |
| 121 | + alarm, auth, and FCM token endpoints. |
| 122 | +- Record each rollout increase with the reason it is acceptable and the signals |
| 123 | + checked. |
| 124 | +- After reaching full rollout, keep one final next-business-day monitoring pass |
| 125 | + before closing the release monitoring note. |
| 126 | + |
| 127 | +## Release Monitoring Note Template |
| 128 | + |
| 129 | +```md |
| 130 | +Release: |
| 131 | +Build: |
| 132 | +Track: |
| 133 | +Current rollout percentage: |
| 134 | +Release owner: |
| 135 | +Backup owner: |
| 136 | +Last checked: |
| 137 | +Next check: |
| 138 | + |
| 139 | +Signals checked: |
| 140 | +- Play review/policy: |
| 141 | +- Android Vitals crashes/ANRs: |
| 142 | +- Ratings/reviews/tester feedback: |
| 143 | +- Firebase/FCM: |
| 144 | +- Backend errors/latency: |
| 145 | +- Team/support reports: |
| 146 | + |
| 147 | +Decision: |
| 148 | +- Continue / pause / rollback / supersede |
| 149 | + |
| 150 | +Notes: |
| 151 | +``` |
0 commit comments