Skip to content

Fix libcurl timeout for large prompts #1421

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
james-martinez wants to merge 2 commits into
base: main
Choose a base branch
from
+19 −10
Draft

Fix libcurl timeout for large prompts #1421

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
017eb1a
Fix libcurl timeout for large prompts by allowing 0 to be infinite ti…
google-labs-jules[bot] Mar 20, 2026
e5ac8f4
Merge branch 'main' into fix-libcurl-timeout-10420877156247985045
james-martinez Mar 21, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions src/cpp/include/lemon/utils/http_client.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -72,19 +72,19 @@ class HttpClient {
static HttpResponse post(const std::string& url,
const std::string& body,
const std::map<std::string, std::string>& headers = {},
long timeout_seconds = 300);
long timeout_seconds = -1);
Copy link
Copy Markdown

@winmutt winmutt Mar 20, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why not add a configuration option per, with global as a default if not set and then a global default of 300s?

Also, under what conditions do we really even want timeouts?

Copy link
Copy Markdown
Contributor Author

@james-martinez james-martinez Mar 21, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

timeouts are configurable for safety and efficiency.


// Multipart form data POST request
static HttpResponse post_multipart(const std::string& url,
const std::vector<MultipartField>& fields,
long timeout_seconds = 300);
long timeout_seconds = -1);

// Streaming POST request (calls callback for each chunk as it arrives)
static HttpResponse post_stream(const std::string& url,
const std::string& body,
StreamCallback stream_callback,
const std::map<std::string, std::string>& headers = {},
long timeout_seconds = 300);
long timeout_seconds = -1);

// Download file to disk with automatic retry and resume support
static DownloadResult download_file(const std::string& url,
Expand Down
11 changes: 10 additions & 1 deletion src/cpp/server/cli_parser.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -58,7 +58,16 @@ static void add_serve_options(CLI::App* serve, ServerConfig& config) {
"Global timeout for HTTP requests, inference, and readiness checks in seconds")
->envname("LEMONADE_GLOBAL_TIMEOUT")
->type_name("SECONDS")
->default_val(config.global_timeout);
->default_val(config.global_timeout)
->transform([](std::string val) {
try {
long t = std::stol(val);
if (t > 0 && t < 300) return std::string("300");
} catch (...) {
// Ignore invalid values, let CLI11 handle them
}
return val;
});

// Multi-model support: Max loaded models per type slot
serve->add_option("--max-loaded-models", config.max_loaded_models,
Expand Down
12 changes: 6 additions & 6 deletions src/cpp/server/utils/http_client.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -120,8 +120,8 @@ HttpResponse HttpClient::post(const std::string& url,
curl_easy_setopt(curl, CURLOPT_POSTFIELDS, body.c_str());
curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, write_callback);
curl_easy_setopt(curl, CURLOPT_WRITEDATA, &response_body);
// Use provided timeout, or fallback to global default (set via --http-timeout)
curl_easy_setopt(curl, CURLOPT_TIMEOUT, timeout_seconds > 0 ? timeout_seconds : default_timeout_seconds_);
// Use provided timeout (0 = infinite), or fallback to global default (set via --global-timeout)
curl_easy_setopt(curl, CURLOPT_TIMEOUT, timeout_seconds >= 0 ? timeout_seconds : default_timeout_seconds_);
curl_easy_setopt(curl, CURLOPT_USERAGENT, "lemon.cpp/1.0");

// Add custom headers
Expand Down Expand Up @@ -182,8 +182,8 @@ HttpResponse HttpClient::post_multipart(const std::string& url,
curl_easy_setopt(curl, CURLOPT_MIMEPOST, mime);
curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, write_callback);
curl_easy_setopt(curl, CURLOPT_WRITEDATA, &response_body);
// Use provided timeout, or fallback to global default (set via --http-timeout)
curl_easy_setopt(curl, CURLOPT_TIMEOUT, timeout_seconds > 0 ? timeout_seconds : default_timeout_seconds_);
// Use provided timeout (0 = infinite), or fallback to global default (set via --global-timeout)
curl_easy_setopt(curl, CURLOPT_TIMEOUT, timeout_seconds >= 0 ? timeout_seconds : default_timeout_seconds_);
curl_easy_setopt(curl, CURLOPT_USERAGENT, "lemon.cpp/1.0");

CURLcode res = curl_easy_perform(curl);
Expand Down Expand Up @@ -258,8 +258,8 @@ HttpResponse HttpClient::post_stream(const std::string& url,
curl_easy_setopt(curl, CURLOPT_POSTFIELDS, body.c_str());
curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, stream_write_callback);
curl_easy_setopt(curl, CURLOPT_WRITEDATA, &callback_data);
// Use provided timeout, or fallback to global default (set via --http-timeout)
curl_easy_setopt(curl, CURLOPT_TIMEOUT, timeout_seconds > 0 ? timeout_seconds : default_timeout_seconds_);
// Use provided timeout (0 = infinite), or fallback to global default (set via --global-timeout)
curl_easy_setopt(curl, CURLOPT_TIMEOUT, timeout_seconds >= 0 ? timeout_seconds : default_timeout_seconds_);
curl_easy_setopt(curl, CURLOPT_USERAGENT, "lemon.cpp/1.0");

// Add custom headers
Expand Down