elasticapm: introduce ELASTIC_APM_SKIP_SERVER_INFO (#2516)

Introduce this option in order to avoid fetch the apm server info to
get its version, off by default.
Calling that endpoint with the aws lambda extension incurs in a
measurable latency on AWS Lambda.
Enabling this issue requires the receiving APM server version to be greater
than 8.7.0.

Option is still in technical preview.

Author: xrmx

docs/reference/configuration.md

@@ -969,6 +969,23 @@ If set to `True`, the agent will intercept the default `sys.excepthook`, which a
 Whether each transaction should have the process arguments attached. Disabled by default to save disk space.
 
 
+### `skip_server_info` [config-skip-server-info]
+
+| Environment | Django/Flask | Default |
+| --- | --- | --- |
+| `ELASTIC_APM_SKIP_SERVER_INFO` | `SKIP_SERVER_INFO` | `False` |
+
+Whether we should skip the server info check to save some latency on constrained environments like AWS Lambda. Disabled by default.
+
+::::{warning}
+This functionality is in technical preview and may be changed or removed in a future release. Elastic will work to fix any issues, but features in technical preview are not subject to the support SLA of official GA features.
+::::
+
+::::{warning}
+This requires sending data to an APM Server newer than 8.7.0 in order to work properly.
+::::
+
+
 ## Django-specific configuration [config-django-specific]
 
 

elasticapm/conf/__init__.py

@@ -743,6 +743,7 @@ class Config(_ConfigBase):
         default=TRACE_CONTINUATION_STRATEGY.CONTINUE,
     )
     include_process_args = _BoolConfigValue("INCLUDE_PROCESS_ARGS", default=False)
+    skip_server_info = _BoolConfigValue("SKIP_SERVER_INFO", default=False)
 
     @property
     def is_recording(self):

elasticapm/transport/http.py

@@ -203,7 +203,11 @@ def _get_cache_control_max_age(self, response_headers):
 
     def _process_queue(self) -> None:
         if not self.client.server_version:
-            self.fetch_server_info()
+            # this is useful on aws lambda environments where this call incurs in unwanted latency
+            if self.client.config.skip_server_info:
+                logger.debug("Skipping to fetch server info")
+            else:
+                self.fetch_server_info()
         super()._process_queue()
 
     def fetch_server_info(self) -> None:

tests/transports/test_urllib3.py

@@ -520,6 +520,23 @@ def test_fetch_server_info_flat_string(waiting_httpserver, caplog, elasticapm_cl
     assert_any_record_contains(caplog.records, "No version key found in server response")
 
 
+def test_skip_server_info(waiting_httpserver, elasticapm_client):
+    elasticapm_client.config.update(version="1", skip_server_info=True)
+    waiting_httpserver.serve_content(code=202, content="", headers={"Location": "http://example.com/foo"})
+    transport = Transport(
+        waiting_httpserver.url, client=elasticapm_client, headers=elasticapm_client._transport._headers
+    )
+    transport.start_thread()
+    try:
+        url = transport.send("x".encode("latin-1"))
+        assert url == "http://example.com/foo"
+    finally:
+        transport.close()
+
+    assert elasticapm_client.server_version is None
+    assert elasticapm_client.check_server_version(gte=(8, 7, 1))
+
+
 def test_close(waiting_httpserver, elasticapm_client):
     elasticapm_client.server_version = (8, 0, 0)  # avoid making server_info request
     waiting_httpserver.serve_content(code=202, content="", headers={"Location": "http://example.com/foo"})