all the changes

cmd/needle-sidecar/eval/rules.py

@@ -0,0 +1,42 @@
+"""Deterministic baseline for the cause categories. Needle has to beat this."""
+import re
+
+PROBE = re.compile(r"(^|/)(\.env|\.git|wp-|wp/|wp$|wordpress|blog/wp|phpmyadmin|cgi-bin|\.aws|\.ssh|xmlrpc|userfiles|@fs/|etc/passwd)", re.I)
+DEP = re.compile(r"dial|connection refused|eof|no such host|upstream|unavailable|broken pipe|reset by peer", re.I)
+CAP = re.compile(r"timeout|deadline exceeded|pool|memory limit|oom|too many|queue full|exhaust", re.I)
+BUG = re.compile(r"panic|nil pointer|index out of range|typeerror|referenceerror|deadlock|unhandled|chunkloaderror|segfault", re.I)
+LIFE = re.compile(r"healthz|readyz|/ping|retry|cancel|keepalive|vital|\bcls\b|ttfb|fcp|lcp|snapshot|replay|unchanged|session\.load", re.I)
+
+
+def fields(text):
+    d = {}
+    for part in text.split(" | "):
+        k, _, v = part.partition(": ")
+        d[k.lower()] = v
+    return d
+
+
+def classify(text):
+    f = fields(text)
+    status = int(f.get("status") or 0)
+    route = f.get("route", "")
+    blob = " ".join([f.get("title", ""), f.get("message", ""), route])
+    if PROBE.search(route):
+        return "scanner_probe"
+    if BUG.search(blob):
+        return "service_defect"
+    if CAP.search(blob):
+        return "capacity_saturation"
+    if DEP.search(blob):
+        return "dependency_unreachable"
+    if status == 401:
+        return "client_credential_issue"
+    if status in (502, 503, 504):
+        return "dependency_unreachable"
+    if status >= 500:
+        return "service_defect"
+    if status == 404:
+        return "stale_reference"
+    if LIFE.search(blob) or status in (200, 202, 204, 301, 302, 304, 308, 499) or (status == 0 and f.get("kind") in ("span", "log")):
+        return "expected_lifecycle"
+    return "needs_review"