fix(writer): update agentless intake to browser-intake/api/v2/spans endpoint [backport 4.11] (#18549)

Backport #18514 to 4.11

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Munir Abdinur <munir.abdinur@datadoghq.com>

Author: 3 people

ddtrace/internal/writer/writer.py

@@ -545,10 +545,38 @@ class AgentlessTraceWriter(HTTPWriter):
     """
 
     HTTP_METHOD = "POST"
-    # Base URL for the agentless trace JSON intake (EvP / track_type:spans).
-    INTAKE_HOST = "public-trace-http-intake.logs"
     # Agentless payloads must be under 15 MB.
     MAX_BUFFER_SIZE = 15 << 20  # 15 MB
+    INTAKE_URLS: dict[str, str] = {
+        "datadoghq.com": "https://public-trace-http-intake.logs.datadoghq.com",
+        "datadoghq.eu": "https://public-trace-http-intake.logs.datadoghq.eu",
+        "us3.datadoghq.com": "https://trace.browser-intake-us3-datadoghq.com",
+        "us5.datadoghq.com": "https://trace.browser-intake-us5-datadoghq.com",
+        "ap1.datadoghq.com": "https://browser-intake-ap1-datadoghq.com",
+        "ap2.datadoghq.com": "https://browser-intake-ap2-datadoghq.com",
+        "uk1.datadoghq.com": "https://browser-intake-uk1-datadoghq.com",
+        "datad0g.com": "https://public-trace-http-intake.logs.datad0g.com",
+    }
+    FALLBACK_INTAKE_URL_TEMPLATE = "https://browser-intake-{}.{}"
+
+    @staticmethod
+    def compute_intake_url(site: str) -> str:
+        url = AgentlessTraceWriter.INTAKE_URLS.get(site)
+        if url is not None:
+            return url
+        # Fallback: strip the TLD, replace remaining dots with dashes, reattach TLD.
+        # e.g. "ddog-gov.com"    -> "browser-intake-ddog-gov.com"
+        #      "us2.ddog-gov.com" -> "browser-intake-us2-ddog-gov.com"
+        prefix, _, tld = site.rpartition(".")
+        url = AgentlessTraceWriter.FALLBACK_INTAKE_URL_TEMPLATE.format(prefix.replace(".", "-"), tld)
+        log.warning(
+            "Datadog site %r is not explicitly supported for agentless tracing. "
+            "Attempting to use %r. To resolve this, upgrade to a newer version of "
+            "ddtrace that supports this site, or disable agentless trace export.",
+            site,
+            url,
+        )
+        return url
 
     def __init__(
         self,
@@ -1112,7 +1140,7 @@ def create_trace_writer(
         return LogWriter()
 
     if agentless:
-        intake_url = "https://{}.{}".format(AgentlessTraceWriter.INTAKE_HOST, config._dd_site)
+        intake_url = AgentlessTraceWriter.compute_intake_url(config._dd_site.lower())
         verify_url(intake_url)
         return AgentlessTraceWriter(
             intake_url=intake_url,

ddtrace/internal/writer/writer_client.py

@@ -40,9 +40,9 @@ def __init__(self, buffer_size, max_payload_size):
 
 
 class AgentlessWriterClient(WriterClientBase):
-    """Client for the agentless JSON span intake (EvP / public-trace-http-intake)."""
+    """Client for the agentless span intake (api/v2/spans)."""
 
-    ENDPOINT = "v1/input"
+    ENDPOINT = "api/v2/spans"
 
     def __init__(self, buffer_size: int, max_payload_size: int) -> None:
         super(AgentlessWriterClient, self).__init__(

releasenotes/notes/agentless-intake-extended-datacenter-support-bb0d7f11b6c15223.yaml

@@ -0,0 +1,6 @@
+---
+fixes:
+  - |
+    LLM Observability: Fixes agentless export dropping data on the ``us3``, ``us5``, ``ap1``,
+    and ``ap2`` Datadog sites. This affected customers on these sites when no Datadog Agent was
+    running or agentless export was explicitly enabled (``DD_LLMOBS_AGENTLESS_ENABLED=1``).

tests/tracer/test_writer.py

@@ -499,7 +499,7 @@ def test_agentless_trace_writer_uses_post():
     assert writer.HTTP_METHOD == "POST"
     assert writer.intake_url == "https://public-trace-http-intake.logs.datadoghq.com"
     assert writer._headers.get("dd-api-key") == "test-api-key"
-    assert writer._clients[0].ENDPOINT == "v1/input"
+    assert writer._clients[0].ENDPOINT == "api/v2/spans"
     assert writer._encoder.content_type == "application/json"
 
 
@@ -541,6 +541,56 @@ def test_agentless_trace_writer_encode_traces():
     writer.flush_queue(raise_exc=True)
 
 
+@pytest.mark.subprocess(
+    env={"_DD_APM_TRACING_AGENTLESS_ENABLED": "1", "DD_API_KEY": "test-key"},
+    parametrize={
+        "DD_SITE": [
+            "datadoghq.com",
+            "datadoghq.eu",
+            "us3.datadoghq.com",
+            "us5.datadoghq.com",
+            "ap1.datadoghq.com",
+            "ap2.datadoghq.com",
+            "uk1.datadoghq.com",
+            "datad0g.com",
+        ]
+    },
+)
+def test_agentless_trace_writer_intake_url():
+    """AgentlessTraceWriter sets the correct intake URL for each DD_SITE."""
+    import os
+
+    from ddtrace.internal.writer.writer import AgentlessTraceWriter
+    from ddtrace.trace import tracer
+
+    site = os.environ["DD_SITE"]
+    writer = tracer._span_aggregator.writer
+    assert isinstance(writer, AgentlessTraceWriter)
+    assert writer.intake_url == AgentlessTraceWriter.INTAKE_URLS[site]
+
+
+@pytest.mark.parametrize("site,expected", list(AgentlessTraceWriter.INTAKE_URLS.items()))
+def test_compute_intake_url_known_sites(site, expected):
+    assert AgentlessTraceWriter.compute_intake_url(site) == expected
+
+
+@pytest.mark.parametrize(
+    "site,expected",
+    [
+        ("ap3.datadoghq.com", "https://browser-intake-ap3-datadoghq.com"),
+        ("ddog-gov.com", "https://browser-intake-ddog-gov.com"),
+        ("us2.ddog-gov.com", "https://browser-intake-us2-ddog-gov.com"),
+    ],
+)
+def test_compute_intake_url_unknown_site_uses_browser_intake_fallback(site, expected):
+    with mock.patch("ddtrace.internal.writer.writer.log") as mock_log:
+        result = AgentlessTraceWriter.compute_intake_url(site)
+
+    assert result == expected
+    mock_log.warning.assert_called_once()
+    assert site in mock_log.warning.call_args[0][1]
+
+
 def test_humansize():
     assert _human_size(0) == "0B"
     assert _human_size(999) == "999B"