zzstoatzz.io
search for standard sites
pub-search.waow.tech
search
zig
blog
atproto
add http.request span + slow receive logging for latency diagnosis
server-side search is 0-2.5ms but TTFB is 1-2.5s for some requests.
the new http.request span wraps the full request lifecycle including
receiveHead(), and logs a warning when receive takes >100ms. this
will reveal whether the delay is in Fly proxy, connection queueing,
or request reading.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
+17
backend/src/server.zig
+17
backend/src/server.zig
···
26
26
var server = http.Server.init(reader.interface(), &writer.interface);
27
27
28
28
while (true) {
29
+
const recv_start = std.time.microTimestamp();
29
30
var request = server.receiveHead() catch |err| {
30
31
if (err != error.HttpConnectionClosing and err != error.EndOfStream) {
31
32
logfire.debug("http receive error: {}", .{err});
32
33
}
33
34
return;
34
35
};
36
+
const recv_us = std.time.microTimestamp() - recv_start;
37
+
const target = request.head.target;
38
+
39
+
// log slow receives — this captures time waiting for the request to arrive
40
+
if (recv_us > 100_000) { // > 100ms
41
+
logfire.warn("http.receive slow: {d}ms | {s}", .{ @divTrunc(recv_us, 1000), target });
42
+
}
43
+
44
+
const req_span = logfire.span("http.request", .{
45
+
.target = target,
46
+
.receive_ms = @divTrunc(recv_us, 1000),
47
+
});
48
+
35
49
handleRequest(&server, &request) catch |err| {
36
50
logfire.err("request error: {}", .{err});
51
+
req_span.end();
37
52
return;
38
53
};
54
+
req_span.end();
55
+
39
56
if (!request.head.keep_alive) return;
40
57
}
41
58
}