vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

server: metrics: fix when no prompt processed

Browse source
This commit is contained in:
Pierrick HYMBERT 2024-03-02 19:11:53 +01:00
parent 9fcfa63a11
commit 61b97915b0
2 changed files with 6 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

3
examples/server/server.cpp
View file

@ -2924,9 +2924,10 @@ int main(int argc, char **argv)
for (const auto& metric_def : metrics_def) {
std::string name = metric_def["name"];
std::string help = metric_def["help"];
auto value = json_value(metric_def, "value", 0);
prometheus << "# HELP llamacpp:" << name << " " << help << "\n"
<< "# TYPE llamacpp:" << name << " " << type << "\n"
<< "llamacpp:" << name << " " << metric_def["value"] << "\n";
<< "llamacpp:" << name << " " << value << "\n";
}
}

2
examples/server/tests/features/steps/steps.py
View file

@ -546,6 +546,8 @@ async def step_prometheus_metrics_exported(context):
assert metrics_response.headers['Content-Type'] == "text/plain; version=0.0.4"
metrics_raw = await metrics_response.text()
metric_exported = False
if context.debug:
print(f"/metrics answer:\n{metrics_raw}\n")
for metric in parser.text_string_to_metric_families(metrics_raw):
match metric.name:
case "llamacpp:kv_cache_usage_ratio":