main - Reduce workers to 8, increase cache to 300, redesign monitoring page

Remove /health endpoint and per-request MEM log; monitoring is now
internal only via background reporter. Root page redesigned with card
grid, colour-coded RSS (green/amber/red), cache progress bars and
server-side rendering with meta-refresh.

Dockerfile

@@ -16,4 +16,4 @@ RUN python -c "from tensorflow.keras.applications import MobileNetV2; MobileNetV
 COPY ./*.py /app/
 
 
-CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860", "--workers", "10"]
+CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860", "--workers", "8"]

main.py

@@ -1,4 +1,5 @@
 from contextlib import asynccontextmanager
+from datetime import datetime
 from typing import List
 from fastapi import FastAPI
 from fastapi.responses import HTMLResponse
@@ -16,42 +17,80 @@ async def lifespan(app: FastAPI):
 
 app = FastAPI(title="Prozis Similarity Service", version="1.0.0", lifespan=lifespan)
 
+_CACHE_MAX = 300
 
-@app.get("/health", include_in_schema=False)
-async def health():
-    return get_all_memory_stats()
+
+def _rss_color(rss_mb: float) -> str:
+    if rss_mb < 800:
+        return "#22c55e"
+    if rss_mb < 1000:
+        return "#f59e0b"
+    return "#ef4444"
 
 
 @app.get("/", include_in_schema=False)
 async def root():
-     return HTMLResponse(content=(
-        "<html><head>"
+    workers = get_all_memory_stats()
+    total_rss_gb = sum(w["rss_mb"] for w in workers) / 1024
+    timestamp = datetime.now().strftime("%H:%M:%S")
+
+    cards = ""
+    for w in workers:
+        color = _rss_color(w["rss_mb"])
+        cache_pct = min(100, w["cache_size"] / _CACHE_MAX * 100)
+        cards += (
+            f"<div class='card' style='--c:{color}'>"
+            f"<div class='pid'>pid {w['pid']}</div>"
+            f"<div class='rss'>{w['rss_mb']}<span class='unit'> MB</span></div>"
+            f"<div class='bar-wrap'>"
+            f"<div class='bar-label'><span>cache</span><span>{w['cache_size']}/{_CACHE_MAX}</span></div>"
+            f"<div class='track'><div class='fill' style='width:{cache_pct:.0f}%'></div></div>"
+            f"</div></div>"
+        )
+
+    html = (
+        "<!DOCTYPE html><html><head>"
+        "<meta charset='UTF-8'>"
+        "<meta http-equiv='refresh' content='10'>"
+        "<title>Prozis Similarity Service</title>"
         "<link rel='preconnect' href='https://fonts.googleapis.com'>"
         "<link rel='preconnect' href='https://fonts.gstatic.com' crossorigin>"
-        "<link href='https://fonts.googleapis.com/css2?family=Barlow+Condensed:ital,wght@1,800&display=swap' rel='stylesheet'>"
-        "</head><body style='margin:0;background:#000;padding:2rem'>"
-        "<h1 style='color:#fff;font-family:\"Barlow Condensed\",sans-serif;font-weight:800;font-style:italic'>"
-        "Prozis Similarity Service"
-        "</h1>"
-        "<table id='mem-stats' style='color:#8f8;font-family:monospace;font-size:0.9rem;border-collapse:collapse'></table>"
-        "<script>"
-        "async function refreshStats(){"
-        "try{"
-        "const r = await fetch('/health');"
-        "const workers = await r.json();"
-        "const rows = workers.map(w => "
-        "`<tr><td style='padding:2px 12px 2px 0'>pid ${w.pid}</td>"
-        "<td style='padding:2px 12px 2px 0'>${w.rss_mb} MB</td>"
-        "<td style='padding:2px 0'>cache ${w.cache_size}</td></tr>`"
-        ").join('');"
-        "document.getElementById('mem-stats').innerHTML = rows;"
-        "}catch(e){}"
-        "}"
-        "refreshStats();"
-        "setInterval(refreshStats, 3000);"
-        "</script>"
+        "<link href='https://fonts.googleapis.com/css2?family=Barlow+Condensed:ital,wght@1,800&family=Inter:wght@400;500&display=swap' rel='stylesheet'>"
+        "<style>"
+        "*{margin:0;padding:0;box-sizing:border-box}"
+        "body{background:#0d0d14;color:#e2e8f0;"
+        "font-family:'Inter',system-ui,sans-serif;min-height:100vh;padding:2rem 2.5rem}"
+        "h1{font-family:'Barlow Condensed',sans-serif;font-weight:800;font-style:italic;"
+        "font-size:2.4rem;color:#fff;letter-spacing:-.01em;line-height:1}"
+        ".meta{margin-top:.5rem;color:#4b5563;font-size:.78rem;font-family:'SF Mono','Fira Code',monospace;letter-spacing:.03em}"
+        ".meta b{color:#6b7280;font-weight:500}"
+        "header{margin-bottom:2rem;padding-bottom:1.5rem;border-bottom:1px solid #1a1a2e}"
+        ".grid{display:grid;grid-template-columns:repeat(auto-fill,minmax(160px,1fr));gap:.875rem}"
+        ".card{background:#111827;border:1px solid #1e2435;border-radius:12px;padding:1.1rem 1.2rem;"
+        "position:relative;overflow:hidden;transition:border-color .15s,transform .15s}"
+        ".card:hover{border-color:#2d3748;transform:translateY(-1px)}"
+        ".card::before{content:'';position:absolute;top:0;left:0;right:0;height:2px;background:var(--c)}"
+        ".pid{font-size:.62rem;color:#374151;font-family:'SF Mono','Fira Code',monospace;"
+        "text-transform:uppercase;letter-spacing:.1em;margin-bottom:.65rem}"
+        ".rss{font-size:1.55rem;font-weight:700;font-family:'SF Mono','Fira Code',monospace;"
+        "color:var(--c);line-height:1;white-space:nowrap}"
+        ".unit{font-size:.65rem;color:#6b7280;font-weight:400}"
+        ".bar-wrap{margin-top:.9rem}"
+        ".bar-label{font-size:.62rem;color:#374151;display:flex;justify-content:space-between;margin-bottom:.3rem}"
+        ".track{background:#1a2235;border-radius:99px;height:3px;overflow:hidden}"
+        ".fill{height:100%;border-radius:99px;background:#3b82f6;opacity:.75}"
+        "</style></head><body>"
+        f"<header>"
+        f"<h1>Prozis Similarity Service</h1>"
+        f"<div class='meta'>"
+        f"<b>{len(workers)}</b> workers &nbsp;&middot;&nbsp; "
+        f"<b>{total_rss_gb:.1f} GB</b> RSS &nbsp;&middot;&nbsp; {timestamp}"
+        f"</div></header>"
+        f"<div class='grid'>{cards}</div>"
         "</body></html>"
-    ))
+    )
+
+    return HTMLResponse(content=html)
 
 
 @app.post("/similarity", summary="Perform images similarity", response_model=List[ResponseModel], tags=["Similarities"])

similarity.py

@@ -35,7 +35,7 @@ _mobilenet_lock = threading.Lock()
 
 _image_cache = OrderedDict()
 _image_cache_lock = threading.Lock()
-_IMAGE_CACHE_MAX = 200
+_IMAGE_CACHE_MAX = 300
 
 _process = psutil.Process()
 
@@ -394,8 +394,4 @@ def check_similarity(images: List[RequestModel], quadrant: bool = True):
             for i in range(1, len(images))
         ]
 
-    memory_stats = get_memory_stats()
-    logging.info(f"[MEM] pid={memory_stats['pid']} rss={memory_stats['rss_mb']}MB "
-                 f"cache_size={memory_stats['cache_size']} images_in_request={len(images) - 1}")
-
     return results