ci: bench: add per slot metric in the commit status

phymbert · Apr 3, 2024 · 64c7534 · 64c7534
1 parent 04e1ce3
commit 64c7534
Showing 1 changed file with 11 additions and 6 deletions.
diff --git a/examples/server/bench/bench.py b/examples/server/bench/bench.py
@@ -16,6 +16,7 @@
 import matplotlib.dates
 import matplotlib.pyplot as plt
 import requests
+from statistics import mean
 
 
 def main(args_in: list[str] | None = None) -> None:
@@ -109,6 +110,7 @@ def main(args_in: list[str] | None = None) -> None:
 
     # Prometheus
     end_time = time.time()
+    prometheus_metrics = {}
     if is_server_listening("0.0.0.0", 9090):
         metrics = ['prompt_tokens_seconds', 'predicted_tokens_seconds',
                    'kv_cache_usage_ratio', 'requests_processing', 'requests_deferred']
@@ -127,6 +129,7 @@ def main(args_in: list[str] | None = None) -> None:
                 values = metric_data['data']['result'][0]['values']
                 timestamps, metric_values = zip(*values)
                 metric_values = [float(value) for value in metric_values]
+                prometheus_metrics[metric] = metric_values
                 timestamps_dt = [datetime.fromtimestamp(int(ts)) for ts in timestamps]
                 plt.figure(figsize=(16, 10), dpi=80)
                 plt.plot(timestamps_dt, metric_values, label=metric)
@@ -178,16 +181,18 @@ def main(args_in: list[str] | None = None) -> None:
     bench_results = {
         "i": iterations,
         "req": {
-            "p90": data['metrics']["http_req_duration"]["p(90)"],
-            "avg": data['metrics']["http_req_duration"]["avg"],
+            "p90": round(data['metrics']["http_req_duration"]["p(90)"], 2),
+            "avg": round(data['metrics']["http_req_duration"]["avg"], 2),
         },
         "pp": {
-            "p90": data['metrics']["llamacpp_prompt_tokens"]["p(90)"],
-            "avg": data['metrics']["llamacpp_prompt_tokens"]["avg"],
+            "p90": round(data['metrics']["llamacpp_prompt_tokens"]["p(90)"], 2),
+            "avg": round(data['metrics']["llamacpp_prompt_tokens"]["avg"], 2),
+            "0": round(mean(prometheus_metrics['prompt_tokens_seconds']), 2),
         },
         "tg": {
-            "p90": data['metrics']["llamacpp_tokens_second"]["p(90)"],
-            "avg": data['metrics']["llamacpp_tokens_second"]["avg"],
+            "p90": round(data['metrics']["llamacpp_tokens_second"]["p(90)"], 2),
+            "avg": round(data['metrics']["llamacpp_tokens_second"]["avg"], 2),
+            "0": round(mean(prometheus_metrics['predicted_tokens_seconds']), 2),
         },
     }
     with open("results.github.env", 'a') as github_env: