Reference

Reference: server.limits

server.limits.body_read_timeout

Maximum gap between request body chunks.

server.limits.h2_max_concurrent_streams

HTTP/2 streams a client may have open at once on one connection.

server.limits.header_read_timeout

HTTP/1: time allowed to receive a complete request head; also bounds keep-alive idle.

server.limits.idle_timeout

Close connections (h1 and h2) with no in-flight request for this long.

server.limits.max_connections

Open client connections across all listeners; further connections wait in the kernel backlog.

server.limits.max_header_bytes

Total size of a request’s headers.

server.limits.max_headers

Number of request headers.

server.limits.max_request_body_bytes

Size of a request body.

server.limits.max_upgraded_connections

Concurrent upgraded (WebSocket) tunnels across all applications.

server.limits.max_uri_bytes

Length of the request target (path and query); longer requests get 414.

server.limits.shutdown_delay

After SIGTERM, keep serving for this long while /readyz answers 503, so load balancers stop sending new requests before the drain starts.

server.limits.shutdown_timeout

Time in-flight requests get to finish when the server stops.

server.limits.tls_handshake_timeout

Time a client has to complete the TLS handshake.

server.limits.upgraded_idle_timeout

Close an upgraded tunnel with no traffic in either direction for this long.

server.limits.write_timeout

Close a connection whose writes make no progress for this long (slow readers).