Minor update

triton-inference-server · Aug 6, 2024 · 6f97f6f · 6f97f6f
1 parent 8eba2f0
commit 6f97f6f
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 3 deletions.
diff --git a/ci/L0_backend_vllm/metrics_test/vllm_metrics_test.py b/ci/L0_backend_vllm/metrics_test/vllm_metrics_test.py
@@ -58,7 +58,7 @@ def get_metrics(self):
         r.raise_for_status()
 
         # Regular expression to match the pattern
-        pattern = r"^(vllm:.*){.*} (\d+)$"
+        pattern = r"^(vllm:[^ {]+)(?:{.*})? ([0-9.-]+)$"
         vllm_dict = {}
 
         # Find all matches in the text
@@ -71,7 +71,7 @@ def get_metrics(self):
         return vllm_dict
 
     def test_vllm_metrics(self):
-        # Supported vLLM metrics
+        # All vLLM metrics from tritonserver
         expected_metrics_dict = {
             "vllm:prompt_tokens_total": 0,
             "vllm:generation_tokens_total": 0,

diff --git a/src/utils/metrics.py b/src/utils/metrics.py
@@ -91,7 +91,8 @@ def _log_counter(self, counter, data: Union[int, float]) -> None:
         Returns:
             None
         """
-        counter.increment(data)
+        if data != 0:
+            counter.increment(data)
 
     def log(self, stats: VllmStats) -> None:
         """Logs tracked stats to triton metrics server every iteration.