Community incident signal · updated from GitHub
Is Claude Code having issues?
The newest completed day is within the recent filing baseline.
180 reports on 2026-08-29 · 264.6/day trailing baseline · 68% of baseline
This is not Anthropic uptime and does not declare a root cause. It compares the newest completed UTC day with the prior 28 days of public issue filings. For confirmed outages, check the official Anthropic status page.
Error classes drawing recent reports
Curated exact strings, ranked by 90-day report growth and activity. Open one for first-response steps and evidence.
-
local · growing
Killed: 9 (SIGKILL / OOM killer)
The local Claude Code process crashed while parsing or executing.
-
local · growing
Segmentation fault (SIGSEGV / SIGABRT)
The local Claude Code process crashed while parsing or executing.
-
network · growing
ECONNRESET: socket hang up
A DNS, proxy, TLS or socket layer interrupted the connection.
-
network · growing
getaddrinfo ENOTFOUND (DNS lookup failed)
A DNS, proxy, TLS or socket layer interrupted the connection.
-
integration · converging
Garbled / flickering terminal output
A local integration or child process did not satisfy Claude Code’s expected contract.
-
api · converging
API Error 400 (invalid_request_error)
The API rejected or could not complete the request.
Latest completed release window
v2.1.241 drew 1750 repo-wide filings in the 7 completed days after publication, versus a baseline of 1894.5 (0.92×).
This measures timing, not causation. Reports may concern older builds or unrelated requests.
How to read this status
- Is this the official Claude status page?
- No. This page measures public GitHub issue-report activity. Use the official Anthropic status page for confirmed service incidents.
- Does elevated reporting prove a release caused a regression?
- No. The signal is temporal correlation only. A busy release, duplicate reports, or a service incident can all increase filing volume.
- What should I do when my exact error is listed?
- Open the error page for conservative first-response steps, affected build evidence, backlog direction, and community workaround leads.