send_json/stress_test utilities: prettier tag-sending

bsmain/management/commands/send_json.py

@@ -15,6 +15,8 @@ from compat.dsn import get_store_url, get_envelope_url, get_header_value
 from bugsink.streams import compress_with_zlib, WBITS_PARAM_FOR_GZIP, WBITS_PARAM_FOR_DEFLATE
 from bugsink.utils import nc_rnd
 
+from ..utils import random_for_RANDOM
+
 
 class Command(BaseCommand):
     help = "Send raw events to a sentry-compatible server; events can be sources from the filesystem or your DB."
@@ -125,6 +127,8 @@ class Command(BaseCommand):
             for tag in options["tag"]:
                 tag = tag[0]  # it's a list of lists... how to prevent this is not immediately clear
                 k, v = tag.split(":", 1)
+
+                v = random_for_RANDOM(k, v)
                 data["tags"][k] = v
 
         if options["valid_only"] and not self.is_valid(data, identifier):

bsmain/management/commands/stress_test.py

@@ -15,18 +15,7 @@ from bugsink.streams import compress_with_zlib, WBITS_PARAM_FOR_GZIP, WBITS_PARA
 from bugsink.utils import nc_rnd
 from issues.utils import get_values
 
-
-def random_postfix():
-    # avoids numbers, because when usedd in the type I imagine numbers may at some point be ignored in the grouping.
-    random_number = nc_rnd.random()
-
-    if random_number < 0.1:
-        # 10% of the time we simply sample from 1M to create a "fat tail".
-        unevenly_distributed_number = int(nc_rnd.random() * 1_000_000)
-    else:
-        unevenly_distributed_number = int(1 / random_number)
-
-    return "".join([chr(ord("A") + int(c)) for c in str(unevenly_distributed_number)])
+from ..utils import random_postfix
 
 
 class Command(BaseCommand):

bsmain/management/utils.py

@@ -0,0 +1,43 @@
+from bugsink.utils import nc_rnd
+
+
+PRETTY_RANDOM_TAGS = {
+    "os": ["Windows", "Linux", "MacOS", "Android", "iOS"],
+    "os.version": ["10", "11", "12", "13", "14", "15", "16", "17", "18", "19", "20"],
+    "cpu": ["x86", "x64", "ARM", "ARM64", "MIPS"],
+    "browser": ["Chrome", "Firefox", "Safari", "Edge", "Opera"],
+    "device": ["Desktop", "Laptop", "Tablet", "Smartphone"],
+    "environment": ["production", "staging", "development", "testing"],
+    "release": ["1.0.0", "1.1.0", "1.2.0", "2.0.0", "2.1.0"],
+    "feature_flag": ["new-ui", "beta-feature", "dark-mode", "performance-improvements"],
+}
+
+
+def random_postfix():
+    # avoids numbers, because when used in the type I imagine numbers may at some point be ignored in the grouping.
+    random_number = nc_rnd.random()
+
+    if random_number < 0.1:
+        # 10% of the time we simply sample from 1M to create a "fat tail".
+        unevenly_distributed_number = int(nc_rnd.random() * 1_000_000)
+    else:
+        unevenly_distributed_number = int(1 / random_number)
+
+    return "".join([chr(ord("A") + int(c)) for c in str(unevenly_distributed_number)])
+
+
+def random_for_RANDOM(k, v):
+    if v != "RANDOM":
+        return v
+
+    if k in PRETTY_RANDOM_TAGS:
+        options = PRETTY_RANDOM_TAGS[k]
+        random_number = nc_rnd.random()
+        unevenly_distributed_number = int(1 / random_number)
+
+        if unevenly_distributed_number > len(options) - 1:
+            return options[0]
+
+        return options[unevenly_distributed_number]
+
+    return "value-" + random_postfix()