add benchmark tool

Makefile

@@ -114,7 +114,14 @@ endif
 endif
 
 ifeq ("$(TEST_BENCHMARK)", "true")
-	TAGS := -bench=. -run=notests -cpu=1,2,4
+	# Check if TESTOPTIONS already contains a specific benchmark filter
+	ifneq (,$(findstring -bench=,$(TESTOPTIONS)))
+		# TESTOPTIONS contains a specific benchmark filter, don't add generic -bench=.
+		TAGS := -run=notests -cpu=1,2,4
+	else
+		# No specific benchmark filter, use generic -bench=. to run all benchmarks
+		TAGS := -bench=. -run=notests -cpu=1,2,4
+	endif
 	TESTS := $(REPOPATH)/test
 endif
 
@@ -145,8 +152,8 @@ ifeq ("$(DEBUG)", "true")
 	DOCKER_RUN_CMD := $(DOCKER_DEBUG_ARGS) $(GOIMAGE) /go/bin/dlv --listen=:$(DEBUG_PORT) --headless=true --api-version=2 --accept-multiclient exec /test_debug.test -- $(TESTOPTIONS)
 	DOCKER_V2_RUN_CMD := $(DOCKER_RUN_CMD)
 else
-    DOCKER_RUN_CMD := $(GOIMAGE) go test -timeout 120m $(GOBUILDTAGSOPT) $(TESTOPTIONS) $(TESTVERBOSEOPTIONS) $(TESTS)
-    DOCKER_V2_RUN_CMD := $(GOV2IMAGE) go test -timeout 120m $(GOBUILDTAGSOPT) $(TESTOPTIONS) $(TESTVERBOSEOPTIONS) -parallel $(TESTV2PARALLEL) ./tests
+    DOCKER_RUN_CMD := $(GOIMAGE) go test -timeout 120m $(GOBUILDTAGSOPT) $(TESTOPTIONS) $(TESTVERBOSEOPTIONS) $(TAGS) $(TESTS)
+    DOCKER_V2_RUN_CMD := $(GOV2IMAGE) go test -timeout 120m $(GOBUILDTAGSOPT) $(TESTOPTIONS) $(TESTVERBOSEOPTIONS) $(TAGS) -parallel $(TESTV2PARALLEL) ./tests
 endif
 
 ifeq ("$(ADD_TIMESTAMP)", "true")
@@ -508,6 +515,20 @@ run-benchmarks-single-vpack-no-auth:
 	@echo "Benchmarks: Single server, Velocypack, no authentication"
 	@${MAKE} TEST_MODE="single" TEST_AUTH="none" TEST_CONTENT_TYPE="vpack" TEST_BENCHMARK="true" __run_tests
 
+run-benchmarks-v2-single-json-no-auth: 
+	@echo "V2 Benchmarks: Single server, JSON no authentication"
+	@${MAKE} TEST_MODE="single" TEST_AUTH="none" TEST_CONTENT_TYPE="json" TEST_BENCHMARK="true" __run_v2_tests
+
+# V1 Cluster benchmarks
+run-benchmarks-cluster-json-no-auth: 
+	@echo "V1 Benchmarks: Cluster server, JSON no authentication"
+	@${MAKE} TEST_MODE="cluster" TEST_AUTH="none" TEST_CONTENT_TYPE="json" TEST_BENCHMARK="true" __run_tests
+
+# V2 Cluster benchmarks
+run-benchmarks-v2-cluster-json-no-auth: 
+	@echo "V2 Benchmarks: Cluster server, JSON no authentication"
+	@${MAKE} TEST_MODE="cluster" TEST_AUTH="none" TEST_CONTENT_TYPE="json" TEST_BENCHMARK="true" __run_v2_tests
+
 ## Lint
 
 .PHONY: tools

README.md

@@ -41,3 +41,149 @@ Key:
 * `+` Features included in the driver may be not present in the ArangoDB API.
   Calls to ArangoDB may result in unexpected responses (404).
 * `-` The ArangoDB version has features that are not supported by the driver.
+
+## Running Benchmark Tests
+
+The go-driver includes comprehensive benchmark tests to measure performance of both V1 and V2 APIs. These benchmarks help compare performance between different driver versions and HTTP protocols.
+
+### Prerequisites
+
+- Docker (for running ArangoDB test instances)
+- Go 1.19+ 
+- Make
+
+### Available Benchmark Tests
+
+#### V1 API Benchmarks
+- **Collection Operations**: `BenchmarkCollectionExists`, `BenchmarkCollection`, `BenchmarkCollections`
+- **Document Operations**: `BenchmarkCreateDocument`, `BenchmarkReadDocument`, `BenchmarkRemoveDocument`
+- **Parallel Operations**: `BenchmarkCreateDocumentParallel`, `BenchmarkReadDocumentParallel`
+- **Comprehensive CRUD**: `BenchmarkComprehensiveDocumentOperations_1K`, `BenchmarkComprehensiveDocumentOperations_10K`
+- **Performance Tests**: `Benchmark_Insert`, `Benchmark_BatchInsert`
+
+#### V2 API Benchmarks
+- **Collection Operations**: `BenchmarkV2CollectionExists`, `BenchmarkV2Collection`, `BenchmarkV2Collections`
+- **Document Operations**: `BenchmarkV2CreateDocument`, `BenchmarkV2ReadDocument`, `BenchmarkV2RemoveDocument`
+- **Parallel Operations**: `BenchmarkV2CreateDocumentParallel`, `BenchmarkV2ReadDocumentParallel`
+- **Comprehensive CRUD**: `BenchmarkV2ComprehensiveDocumentOperations_1K`, `BenchmarkV2ComprehensiveDocumentOperations_10K`
+- **Performance Tests**: `Benchmark_Insert`, `Benchmark_BatchInsert`
+
+### Running Benchmarks
+
+#### Basic Benchmark Execution
+
+```bash
+# Run all V1 benchmarks
+make run-benchmarks-single-json-no-auth
+
+# Run all V2 benchmarks  
+make run-benchmarks-v2-single-json-no-auth
+```
+
+#### Running Specific Benchmarks
+
+```bash
+# V1 API - Specific benchmark
+export TESTOPTIONS="-bench=BenchmarkCreateDocument -test.v"
+make run-benchmarks-single-json-no-auth
+
+# V2 API - Specific benchmark
+export TESTOPTIONS="-bench=BenchmarkV2CreateDocument -test.v"
+make run-benchmarks-v2-single-json-no-auth
+
+# V1 API - Comprehensive CRUD with 10K documents
+export TESTOPTIONS="-bench=BenchmarkComprehensiveDocumentOperations_10K -test.v"
+make run-benchmarks-single-json-no-auth
+
+# V2 API - Comprehensive CRUD with 10K documents
+export TESTOPTIONS="-bench=BenchmarkV2ComprehensiveDocumentOperations_10K -test.v"
+make run-benchmarks-v2-single-json-no-auth
+```
+
+#### Running Multiple Benchmarks
+
+```bash
+# Run multiple V2 benchmarks
+export TESTOPTIONS="-bench=BenchmarkV2CreateDocument|BenchmarkV2ReadDocument -test.v"
+make run-benchmarks-v2-single-json-no-auth
+
+# Run both 1K and 10K comprehensive benchmarks
+export TESTOPTIONS="-bench=BenchmarkV2ComprehensiveDocumentOperations_1K|BenchmarkV2ComprehensiveDocumentOperations_10K -test.v"
+make run-benchmarks-v2-single-json-no-auth
+```
+
+#### Comparing V1 vs V2 Performance
+
+```bash
+# Run V1 10K comprehensive benchmark
+export TESTOPTIONS="-bench=BenchmarkComprehensiveDocumentOperations_10K -test.v"
+make run-benchmarks-single-json-no-auth
+
+# Run V2 10K comprehensive benchmark
+export TESTOPTIONS="-bench=BenchmarkV2ComprehensiveDocumentOperations_10K -test.v"
+make run-benchmarks-v2-single-json-no-auth
+```
+
+### Benchmark Output Explanation
+
+The benchmark output includes:
+
+- **Operations per second**: How many operations completed in the benchmark duration
+- **Time per operation**: Average time per operation (e.g., `720955 ns/op`)
+- **Memory allocations**: Bytes allocated per operation (e.g., `9505 B/op`)
+- **Allocation count**: Number of memory allocations per operation (e.g., `114 allocs/op`)
+- **CPU scaling**: Results for different CPU counts (1, 2, 4 cores)
+
+Example output:
+```
+BenchmarkV2CreateDocument/HTTP_JSON-4    1197    1055266 ns/op    10713 B/op    128 allocs/op
+```
+
+This means:
+- 1197 operations completed
+- 1,055,266 nanoseconds per operation (~1.06ms)
+- 10,713 bytes allocated per operation
+- 128 memory allocations per operation
+- Tested with 4 CPU cores
+
+### HTTP Protocol Comparison
+
+V2 benchmarks automatically test both HTTP/1.1 and HTTP/2 protocols:
+
+```
+BenchmarkV2CreateDocument/HTTP_JSON      # HTTP/1.1 results
+BenchmarkV2CreateDocument/HTTP2_JSON     # HTTP/2 results
+```
+
+### Benchmark Options
+
+You can customize benchmark execution with additional options:
+
+```bash
+# Set benchmark duration
+export TESTOPTIONS="-bench=BenchmarkV2CreateDocument -benchtime=10s -test.v"
+
+# Set number of iterations
+export TESTOPTIONS="-bench=BenchmarkV2CreateDocument -count=5 -test.v"
+
+# Run with memory profiling
+export TESTOPTIONS="-bench=BenchmarkV2CreateDocument -memprofile=mem.prof -test.v"
+
+# Run with CPU profiling  
+export TESTOPTIONS="-bench=BenchmarkV2CreateDocument -cpuprofile=cpu.prof -test.v"
+```
+
+### Performance Tips
+
+1. **Use specific benchmark names** to avoid running all tests
+2. **Compare HTTP/1.1 vs HTTP/2** performance in V2 benchmarks
+3. **Test with different document counts** (1K vs 10K) to understand scaling
+4. **Run multiple iterations** for more reliable results
+5. **Use profiling options** to identify performance bottlenecks
+
+### Troubleshooting
+
+- **"No endpoints found"**: Ensure Docker is running and the test environment is properly set up
+- **Benchmark not found**: Check the exact function name (case-sensitive)
+- **All benchmarks running**: Use specific benchmark names in `TESTOPTIONS`
+- **HTTP/2 errors**: V2 benchmarks automatically handle HTTP/2 stream limits