Network path, timeout budget, and connection management issues.
Last reviewed: February 28, 2026|8 mapped errors
Network-timeout errors emerge when transport path issues or mismatched timeout budgets prevent requests from completing in expected windows.
Provider-specific error pages mapped to this category (8 total). Showing 1-8.
GCP DEADLINE_EXCEEDED means the client deadline expired before receiving a response, even if server-side work might have completed.
HTTP 101 Switching Protocols is an informational success response indicating that the server agrees to switch the application protocol of the current connection. It is the mandatory handshake response for establishing WebSockets (wss://) or upgrading to HTTP/2.
HTTP 103 Early Hints is an informational response sent by the server before the final status code. It contains Link headers that allow browsers to begin preloading critical resources-like CSS, JavaScript, and Fonts-while the server is still processing the main request.
HTTP 203 Non-Authoritative Information is a success status code indicating that the request was successful, but the enclosed payload has been modified by a transforming proxy or API gateway from the origin server's original 200 OK response.
HTTP 226 IM Used is a success status code indicating that the server has fulfilled a GET request by returning a "delta" (a set of changes) rather than the full resource. It requires the client to apply these instance manipulations to a previously cached base version.
HTTP 408 Request Timeout means the server did not receive a complete request message within its timeout window.
HTTP 421 Misdirected Request means the request reached a server that is not able to produce a response for the target authority.
HTTP 426 Upgrade Required means the server refuses current protocol and requires the client to switch protocols.
Because upstream gateways or dependencies can still expire first. Timeout budgets must be aligned across every hop, not only client-side.
Deadline inversion happens when an upstream component has a longer timeout than its caller, allowing work to continue after caller timeout and wasting capacity.
Use hop-level telemetry: DNS/TLS/connection metrics identify transport failures, while application spans show server processing and dependency delay.