New AWS Lambda Support (instana#219)

* Refactor, reorganize and cleanup for AWS prep

* AWS Lambda Agent, Layer, Span & handler method

* Functional Lambda Tracing and Infra linking

* Bug fixes for test suite issues

* Bug fixes for test suite issues: second batch

* Restructured tag collection

* Update suds and grpcio tests.  Unify spans in span.py

* Remove debug remnants

* Update/fix cassandra tests

* Check if agent supports extra_headers before proceeding

* Centralize all lambda instrumentation

* New Lambda test suite

* New and improved lambda handler and handler parsing

* Fix trailing slash bug reported by Justyn

* Refinement and additions for the test suite

* Linter love

* Allow get/set of agent & tracer for tests

* Updated build script

* Restore agent and tracer after each test; lint fixes

* Enable script for all regions

* Better script name

* Trigger support & tests

* Move Lambda inst into it's own package

* Code documentation

* Update import path

* Keep 2.7 compatible: no type hints

* Py 2.7 compatible decompression

* Fix Python 3.5 compatibility

Author: pglombardo

bin/lambda_build_publish_layer.py

@@ -0,0 +1,97 @@
+#!/usr/bin/env python
+
+import os
+import json
+import shutil
+import time
+import distutils.spawn
+from subprocess import call, check_output
+
+# Disable aws CLI pagination
+os.environ["AWS_PAGER"] = ""
+
+# Check requirements first
+for cmd in ["pip", "zip"]:
+    if distutils.spawn.find_executable(cmd) is None:
+        print("Can't find required tool: %s" % cmd)
+        exit(1)
+
+# Determine where this script is running from
+this_file_path = os.path.dirname(os.path.realpath(__file__))
+
+# Change directory to the base of the Python sensor repository
+os.chdir(this_file_path + "/../")
+
+cwd = os.getcwd()
+print("===> Working directory is: %s" % cwd)
+
+# For development, respect or set PYTHONPATH to this repository
+local_env = os.environ.copy()
+if "PYTHONPATH" not in os.environ:
+    local_env["PYTHONPATH"] = os.getcwd()
+
+build_directory = os.getcwd() + '/build/lambda/python'
+
+if os.path.isdir(build_directory):
+    print("===> Cleaning build pre-existing directory: %s" % build_directory)
+    shutil.rmtree(build_directory)
+
+print("===> Creating new build directory: %s" % build_directory)
+os.makedirs(build_directory, exist_ok=True)
+
+print("===> Installing Instana and dependencies into build directory")
+call(["pip", "install", "-q", "-U", "-t", os.getcwd() + '/build/lambda/python', "instana"], env=local_env)
+
+print("===> Manually copying in local dev code")
+shutil.rmtree(build_directory + "/instana")
+shutil.copytree(os.getcwd() + '/instana', build_directory + "/instana")
+
+print("===> Creating Lambda ZIP file")
+timestamp = time.strftime("%Y-%m-%d_%H:%M:%S")
+zip_filename = "instana-py-layer-%s.zip" % timestamp
+
+os.chdir(os.getcwd() + "/build/lambda/")
+call(["zip", "-q", "-r", zip_filename, "./python", "-x", "*.pyc", "./python/pip*", "./python/setuptools*", "./python/wheel*"])
+
+fq_zip_filename = os.getcwd() + '/%s' % zip_filename
+aws_zip_filename = "fileb://%s" % fq_zip_filename
+print("Zipfile should be at: ", fq_zip_filename)
+
+regions = ['ap-northeast-1', 'ap-northeast-2', 'ap-south-1', 'ap-southeast-1', 'ap-southeast-2', 'ca-central-1',
+           'eu-central-1', 'eu-north-1', 'eu-west-1', 'eu-west-2', 'eu-west-3', 'sa-east-1', 'us-east-1',
+           'us-east-2', 'us-west-1', 'us-west-2']
+
+# regions = ['us-west-1']
+
+# LAYER_NAME = "instana-py-test"
+LAYER_NAME = "instana-python"
+
+published = dict()
+
+for region in regions:
+    print("===> Uploading layer to AWS %s " % region)
+    response = check_output(["aws", "--region", region, "lambda", "publish-layer-version",
+                             "--description",
+                             "Provides Instana tracing and monitoring of AWS Lambda functions built with Python",
+                             "--license-info", "MIT", "--output", "json",
+                             "--layer-name", LAYER_NAME, "--zip-file", aws_zip_filename,
+                             "--compatible-runtimes", "python2.7", "python3.6", "python3.7", "python3.8"])
+
+    json_data = json.loads(response)
+    version = json_data['Version']
+    print("===> Uploaded version is %s" % version)
+
+    print("===> Making layer public...")
+    response = check_output(["aws", "--region", region, "lambda", "add-layer-version-permission",
+                             "--layer-name", LAYER_NAME, "--version-number", str(version),
+                             "--statement-id", "public-permission-all-accounts",
+                             "--principal", "*",
+                             "--action", "lambda:GetLayerVersion",
+                             "--output", "text"])
+
+    published[region] = json_data['LayerVersionArn']
+
+
+print("===> Published list:")
+for key in published.keys():
+    print("%s\t%s" % (key, published[key]))

docker-compose.yml

@@ -40,20 +40,37 @@ services:
       - ./bin:/nodejs-collector-bin
 #    command: ["/nodejs-collector-bin/wait-for-it.sh", "-s", "-t", "120", "zookeeper:2181", "--", "start-kafka.sh"]
 
-  mysql:
-    image: mysql:8.0.1
+  cassandra:
+    image: cassandra:3.11.5
+    ports:
+      - 9042:9042
+
+
+  couchbase:
+    image: couchbase
+    ports:
+      - 8091-8094:8091-8094
+      - 11210:11210
+
+  mariadb:
+    image: mariadb
     ports:
       - 3306:3306
     environment:
-      MYSQL_ALLOW_EMPTY_PASSWORD: 'true'
+      MYSQL_DATABASE: 'circle_test'
+      MYSQL_USER: 'root'
+      MYSQL_PASSWORD: ''
+      MYSQL_ALLOW_EMPTY_PASSWORD: 'yes'
       MYSQL_ROOT_PASSWORD: ''
-      MYSQL_DATABASE: circle_test
-      MYSQL_USER: root
-      MYSQL_PASSWORD:
-      MYSQL_ROOT_HOST: '0.0.0.0'
+      MYSQL_ROOT_HOST: '%'
     volumes:
       - ./tests/config/database/mysql/conf.d:/etc/mysql/conf.d
 
+  mongodb:
+    image: 'mongo:3.4.1'
+    ports:
+      - '27017:27017'
+
   postgres:
     image: postgres:10.5
     ports:
@@ -68,3 +85,6 @@ services:
     ports:
       - 5671:5671
       - 5672:5672
+
+#volumes:
+#  mysql-data:

instana/__init__.py

@@ -21,8 +21,9 @@
 
 import os
 import sys
-from threading import Timer
+import importlib
 import pkg_resources
+from threading import Timer
 
 __author__ = 'Instana Inc.'
 __copyright__ = 'Copyright 2020 Instana Inc.'
@@ -45,6 +46,57 @@ def load(_):
     if "INSTANA_DEBUG" in os.environ:
         print("Instana: activated via AUTOWRAPT_BOOTSTRAP")
 
+    if "INSTANA_ENDPOINT_URL" in os.environ:
+        print("load: detected lambda environment")
+
+
+def get_lambda_handler_or_default():
+    """
+    For instrumenting AWS Lambda, users specify their original lambda handler in the LAMBDA_HANDLER environment
+    variable.  This function searches for and parses that environment variable or returns the defaults.
+
+    The default handler value for AWS Lambda is 'lambda_function.lambda_handler' which
+    equates to the function "lambda_handler in a file named "lambda_function.py" or in Python
+    terms "from lambda_function import lambda_handler"
+    """
+    handler_module = "lambda_function"
+    handler_function = "lambda_handler"
+
+    try:
+        handler = os.environ.get("LAMBDA_HANDLER", False)
+
+        if handler:
+            parts = handler.split(".")
+            handler_function = parts.pop()
+            handler_module = ".".join(parts)
+    except:
+        pass
+
+    return handler_module, handler_function
+
+
+def lambda_handler(event, context):
+    """
+    Entry point for AWS Lambda monitoring.
+
+    This function will trigger the initialization of Instana monitoring and then call
+    the original user specified lambda handler function.
+    """
+    module_name, function_name = get_lambda_handler_or_default()
+
+    try:
+        # Import the module specified in module_name
+        handler_module = importlib.import_module(module_name)
+    except ImportError:
+        print("Couldn't determine and locate default module handler: %s.%s", module_name, function_name)
+    else:
+        # Now get the function and execute it
+        if hasattr(handler_module, function_name):
+            handler_function = getattr(handler_module, function_name)
+            return handler_function(event, context)
+        else:
+            print("Couldn't determine and locate default function handler: %s.%s", module_name, function_name)
+
 
 def boot_agent():
     """Initialize the Instana agent and conditionally load auto-instrumentation."""
@@ -56,6 +108,8 @@ def boot_agent():
     # Instrumentation
     if "INSTANA_DISABLE_AUTO_INSTR" not in os.environ:
         # Import & initialize instrumentation
+        from .instrumentation.aws import lambda_inst
+
         if sys.version_info >= (3, 5, 3):
             from .instrumentation import asyncio
             from .instrumentation.aiohttp import client