Eas Update Insights

EAS service - costs apply. Insights cover updates published through EAS Update, a paid Expo Application Services product with free-tier limits. Update delivery and the data behind these commands count against your plan’s EAS Update usage. Review https://expo.dev/pricing (opens in a new tab).

Query the health of published EAS Update directly from the CLI: launches, failed launches, crash rates, unique users, payload size, the embedded-vs-OTA user split per channel, and the most popular updates per runtime version. The data is the same data that powers the update and channel detail pages on expo.dev; these commands expose it in the terminal in human and JSON form.

When to use this skill

Use this when the user wants to assess the health or adoption of a published EAS Update: crash rates, install counts, unique users, bundle size, or the split between embedded and OTA users on a channel.

Example prompts:

• “How is the latest update doing?”
• “Is the latest update healthy?”
• “Is the new release crashing more than the last one?”
• “How many users are on the latest update vs the embedded build?”
• “Which update is most popular on production right now?”
• “How big is our update bundle?”

Also fits: post-publish rollout monitoring and regression detection.

Don’t use when the user needs per-user crash detail or device-level reporting; this skill only exposes aggregate EAS metrics.

Prerequisites

• eas-cli installed (npm install -g eas-cli).
• Logged in: eas login.
• For channel:insights: run from an Expo project directory (the command resolves the project ID from app.json). update:insights only needs a login.

Commands at a glance

All of these support --json --non-interactive for programmatic parsing.

Discovering IDs

Before querying insights for an update group, you need its group ID. Use eas update:list with either --branch <name> (updates on that branch) or --all (updates across all branches). Always pass --json --non-interactive when running non-interactively; without a branch/--all flag the command will otherwise prompt for a branch selection:

# Latest group id across all branches
eas update:list --all --json --non-interactive | jq -r '.currentPage[0].group'

# Latest group id on a specific branch
eas update:list --branch production --json --non-interactive | jq -r '.currentPage[0].group'

The JSON response has a currentPage array with one entry per update group (both platforms of the same publish are collapsed into one entry):

{
  "currentPage": [
    {
      "branch": "production",
      "message": "\"Fix checkout crash\" (1 week ago by someone)",
      "runtimeVersion": "1.0.6",
      "group": "03d5dfcf-736c-475a-8730-af039c3f4d06",
      "platforms": "android, ios",
      "isRollBackToEmbedded": false
    }
  ]
}

Entries also carry codeSigningKey and rolloutPercentage, but only when those features are in use for the group (undefined values are omitted from the JSON output).

When called with --branch <name>, the response also includes name (the branch name) and id (the branch ID) at the top level.

eas update:insights <groupId>

Shows launches, failed launches, crash rate, unique users, launch asset count, and average payload size for a single update group, broken down per platform (iOS, Android), plus a daily breakdown of launches and failures.

Basic use

eas update:insights 03d5dfcf-736c-475a-8730-af039c3f4d06

Flags

JSON output shape

Top level: groupId, timespan (start, end, daysBack), and platforms[] with one entry per platform the group was published to. Each platform entry has updateId, totals (uniqueUsers, installs, failedInstalls, crashRatePercent), payload (launchAssetCount, averageUpdatePayloadBytes), and a daily[] time series of { date, installs, failedInstalls }.

For the complete schema and field reference, see references/update-insights-schema.md (opens in a new tab).

Fields that matter for health assessment:

• platforms[].totals.crashRatePercent, computed as failedInstalls / (installs + failedInstalls) * 100. Zero when there are no installs.
• platforms[].totals.installs and uniqueUsers give the adoption signal.
• platforms[].daily is a time series, useful for spotting a sudden spike in failures.

Errors

• Could not find any updates with group ID: "<id>" — group doesn’t exist or you lack access.
• Update group "<id>" has no ios update (available platforms: android) — --platform ios was used but the group wasn’t published for iOS.
• EAS Update insights is not supported by this version of eas-cli. Please upgrade ... — the server deprecated a field the CLI relies on. Run npm install -g eas-cli@latest.

eas update:view <groupId> --insights

Extends the standard update:view output with the same per-platform insights, inline.

# Human-readable
eas update:view 03d5dfcf-... --insights
eas update:view 03d5dfcf-... --insights --days 30

# JSON: wrapped as { updates: [...], insights: {...} }
eas update:view 03d5dfcf-... --json --insights

Without --insights, update:view behaves exactly as before — no JSON shape change for existing consumers. The --days / --start / --end flags only apply when --insights is set; passing them alone errors.

eas channel:insights --channel <name> --runtime-version <version>

Shows, per channel, how many users are on the embedded build vs over-the-air updates and which updates are pulling the most traffic. Must be run from an Expo project directory.

Basic use

eas channel:insights --channel production --runtime-version 1.0.6

Flags

JSON output shape

Top level: channel, runtimeVersion, timespan, embeddedUpdateTotalUniqueUsers, otaTotalUniqueUsers, mostPopularUpdates[] (each with rank, groupId, message, platform, totalUniqueUsers), cumulativeMetricsAtLastTimestamp[], plus chart-shaped uniqueUsersOverTime and cumulativeMetricsOverTime objects with labels and datasets.

For the complete schema and field reference, see references/channel-insights-schema.md (opens in a new tab).

Fields that matter:

• embeddedUpdateTotalUniqueUsers is the count of users running the embedded (binary-bundled) build.
• mostPopularUpdates[] is updates ranked by totalUniqueUsers. Caveat: this is the top-N the server returns; otaTotalUniqueUsers is a sum of that list and may undercount total OTA reach if more than top-N updates are active.
• uniqueUsersOverTime and cumulativeMetricsOverTime are daily data series for charting.

Errors

• Could not find channel with the name <name> — typo or wrong account.
• “No update launches recorded” in the table / empty mostPopularUpdates in JSON — no OTA update has been launched for that channel + runtime yet. Usually means the channel is still serving the embedded build only.

Common workflows

Verify the update I just published is healthy

# 1. Grab the latest publish on production
GROUP_ID=$(eas update:list --branch production --json --non-interactive \
  | jq -r '.currentPage[0].group')

# 2. Give it some adoption time (minutes to hours), then check crash rate
eas update:insights "$GROUP_ID" --json --non-interactive \
  | jq '.platforms[] | {platform, installs: .totals.installs, crashRate: .totals.crashRatePercent}'

Compare the crashRate across platforms and against previous releases; sudden spikes or asymmetric behaviour (iOS spiking while Android is flat, or vice versa) is the signal to investigate.

Compare adoption between two channels

for channel in production staging; do
  echo "--- $channel ---"
  eas channel:insights --channel "$channel" --runtime-version 1.0.6 --json --non-interactive \
    | jq '{
        channel,
        embedded: .embeddedUpdateTotalUniqueUsers,
        ota: .otaTotalUniqueUsers,
        topUpdate: .mostPopularUpdates[0]
      }'
done

Detect a rollout regression in the last 24 hours

eas update:insights "$GROUP_ID" --days 1 --json --non-interactive \
  | jq '.platforms[] | select(.totals.crashRatePercent > 1)'

Summarize group metrics for release notes

eas update:view "$GROUP_ID" --insights --days 30

Human-readable group details plus 30 days of launches/failures per platform — suitable for pasting into a changelog or incident review.

Output tips

• Pipe JSON through jq; payloads are structured for easy filtering.
• --json implies --non-interactive, but passing both is explicit and scripting-friendly.
• Dates in daily[].date are UTC ISO timestamps; the human-readable table renders them as YYYY-MM-DD (UTC).
• The CLI table labels say “Launches” / “Crashes” while JSON uses installs / failedInstalls. Same field, different display name.

Limitations

• Unique users across platforms may double-count users who run the same publish on both iOS and Android. The same caveat applies to otaTotalUniqueUsers in channel insights, which is a sum over mostPopularUpdates.
• Fresh publishes may show zeros for a short period while the metrics pipeline catches up.
• Installs are downloads, not launches: the installs / "Launches" field counts users who downloaded the manifest and launch asset. A confirmed run only registers on the user's next update check (typically up to 24h later, depending on the app’s update policy). So metrics lag the real-world state slightly.
• Crashes are self-reported: failedInstalls / "Crashes" counts updates that errored during install/launch and were reported on the next update check. Crashes that don't trigger an update request (e.g. process kill before recovery) won't appear.

Submitting Feedback

If you encounter errors, misleading or outdated information in this skill, report it so Expo can improve:

npx --yes submit-expo-feedback@latest --category skills --subject "eas-update-insights" "<actionable feedback>"

Only submit when you have something specific and actionable to report. Include as much relevant context as possible. If an AI agent repeatedly failed or the user had to take over an Expo task, load the expo-skill-feedback skill and follow its eval-candidate flow instead of reusing the command above.