[sentry-junior]

Fixes the messaging and telemetry behind CLI-29S ("Failed to create release: Network error").

unwrapResult re-threw only AuthError/ApiError from the fetch layer. throwApiError then wrapped everything else as Network error … Check your internet connection and try again. Three problems came from that:

• Our own TimeoutError lost its guidance. A request timeout already carries a hint that points at Sentry's status, but the wrapper replaced it with "check your internet connection". Any CliError thrown by the fetch layer now passes through unchanged (this also covers HostScopeError).
• The real cause was dropped. For a connection failure, undici rejects with TypeError: fetch failed and puts the errno (ENOTFOUND, ECONNREFUSED, …) on cause, which was never shown. The detail now names the Sentry URL, the errno message, and a next step for that errno. It ends with sentry status on SaaS, or a SENTRY_URL/SENTRY_HOST check on self-hosted.
• The silencing added for CLI-16W was bypassed. It only matched the raw TypeError, and SDK-backed calls never surface the raw error. The wrapper now keeps the original on cause, and classifySilenced treats a wrapped fetch failed as network_error. TLS errors (ApiError(0) without that cause) and non-network causes, such as a local SQLite unable to open database file, stay captured.

A review question: silencing means ECONNRESET-style failures against sentry.io show up only in the silenced-error metric, not as issues. Timeouts stay captured as TimeoutError.

Before / after

Real sentry release create runs from source, on the base commit and on this branch. For the demo the request timeout was cut to 2s (it is 30s), and the sentry.io case was simulated by routing through a local port that never answers. [exit N] is the exit code.

Before: every case printed the same thing. Connection failures showed Cause: fetch failed. Timeouts showed Cause: Request timed out after 2s., with the same "check your internet connection" advice and exit code 30.

$ SENTRY_URL=https://sentry.acme-corp.invalid sentry release create my-org/1.0.0
Error: Failed to create release: Network error
  Unable to reach Sentry API. Cause: fetch failed

  Check your internet connection and try again.
[exit 30]

After: each message names the server, the real cause, and the next step that fits it, in the CLI's existing error formats:

• Network errors keep the Network error title (so grouping is unchanged) and use the ApiError detail layout from the 401/403 enrichment: an indented detail, a blank line, then one step per line, with commands as Verb with: <command>.
• Timeouts keep TimeoutError's layout and exit code (31): a sentence, then an indented action line, like the View in Sentry:/Or retry: lines in the issue explain and issue plan timeout hints.
• The check step is the CLI's own sentry status command. For a self-hosted timeout it's sentry status --url <server>, which probes /_health/. The login hint uses getUrlLoginCommand(), like the other login hints.

Self-hosted, hostname doesn't resolve:

$ SENTRY_URL=https://sentry.acme-corp.invalid sentry release create my-org/1.0.0
Error: Failed to create release: Network error
  Unable to reach Sentry at https://sentry.acme-corp.invalid: getaddrinfo ENOTFOUND sentry.acme-corp.invalid

  The hostname could not be resolved. Check that the URL is right and that DNS works on this machine.
  This URL comes from SENTRY_URL. If it is not your Sentry server, set SENTRY_URL to the right URL.
[exit 30]

Self-hosted, server not running:

$ SENTRY_URL=http://127.0.0.1:59999 sentry release create my-org/1.0.0
Error: Failed to create release: Network error
  Unable to reach Sentry at http://127.0.0.1:59999: connect ECONNREFUSED 127.0.0.1:59999

  The server refused the connection. Check that the URL and port are right and that Sentry is running.
  This URL comes from SENTRY_URL. If it is not your Sentry server, set SENTRY_URL to the right URL.
[exit 30]

Self-hosted, server accepts the connection but never answers:

$ SENTRY_URL=http://127.0.0.1:8765 sentry release create my-org/1.0.0
Error: Failed to create release: Request timed out after 2s.

The Sentry API at http://127.0.0.1:8765 did not respond in time. The change may still have been applied; check before trying again.
  Check the server's health with: sentry status --url http://127.0.0.1:8765
[exit 31]

sentry.io never answers:

$ sentry release create my-org/1.0.0
Error: Failed to create release: Request timed out after 2s.

The Sentry API at https://sentry.io did not respond in time. The change may still have been applied; check before trying again.
  Check Sentry's status with: sentry status
[exit 31]

The CLI's API requests use plain Node fetch, which ignores HTTPS_PROXY/HTTP_PROXY unless NODE_USE_ENV_PROXY is set. So these messages don't name a proxy, and a proxy variable never replaces the hint for the actual error.

via bete.

--

View Session [Sentry]

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated
sentry-local	Skipped		Nov 4, 2026 9:56am UTC

[cursor]

Stale Bugbot comment from a previous run.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit d3771ca. Configure here.}