|
| 1 | +package com.google.cloud.compute.v1.integration; |
| 2 | + |
| 3 | +import static com.google.common.truth.Truth.assertThat; |
| 4 | +import static org.junit.Assert.fail; |
| 5 | + |
| 6 | +import com.google.api.gax.tracing.CompositeTracerFactory; |
| 7 | +import com.google.api.gax.tracing.OpenTelemetryTracingFactory; |
| 8 | +import com.google.api.gax.tracing.ObservabilityAttributes; |
| 9 | +import com.google.api.gax.tracing.OpenTelemetryMetricsFactory; |
| 10 | +import com.google.api.gax.tracing.LoggingTracerFactory; |
| 11 | +import com.google.api.gax.tracing.ApiTracerFactory; |
| 12 | +import com.google.cloud.compute.v1.InstancesClient; |
| 13 | +import com.google.cloud.compute.v1.InstancesSettings; |
| 14 | +import com.google.cloud.trace.v1.TraceServiceClient; |
| 15 | +import com.google.cloud.trace.v1.TraceServiceSettings; |
| 16 | +import com.google.devtools.cloudtrace.v1.GetTraceRequest; |
| 17 | +import com.google.api.gax.retrying.RetrySettings; |
| 18 | +import com.google.api.gax.rpc.StatusCode; |
| 19 | +import java.time.Duration; |
| 20 | +import com.google.devtools.cloudtrace.v1.Trace; |
| 21 | +import com.google.devtools.cloudtrace.v1.TraceSpan; |
| 22 | +import io.opentelemetry.api.OpenTelemetry; |
| 23 | +import io.opentelemetry.api.trace.Span; |
| 24 | +import io.opentelemetry.api.trace.SpanContext; |
| 25 | +import io.opentelemetry.api.trace.TraceFlags; |
| 26 | +import io.opentelemetry.api.trace.TraceState; |
| 27 | +import io.opentelemetry.api.trace.Tracer; |
| 28 | +import io.opentelemetry.context.Context; |
| 29 | +import io.opentelemetry.context.Scope; |
| 30 | +import io.opentelemetry.sdk.OpenTelemetrySdk; |
| 31 | +import io.opentelemetry.sdk.trace.SdkTracerProvider; |
| 32 | +import io.opentelemetry.sdk.trace.export.BatchSpanProcessor; |
| 33 | +import io.opentelemetry.exporter.otlp.trace.OtlpGrpcSpanExporter; |
| 34 | +import io.opentelemetry.sdk.testing.exporter.InMemoryMetricReader; |
| 35 | +import io.opentelemetry.sdk.metrics.SdkMeterProvider; |
| 36 | +import io.opentelemetry.sdk.metrics.data.MetricData; |
| 37 | +import java.util.Collection; |
| 38 | +import com.google.auth.oauth2.GoogleCredentials; |
| 39 | +import org.slf4j.LoggerFactory; |
| 40 | +import ch.qos.logback.classic.Logger; |
| 41 | +import ch.qos.logback.classic.spi.ILoggingEvent; |
| 42 | +import ch.qos.logback.core.AppenderBase; |
| 43 | +import java.util.ArrayList; |
| 44 | +import java.util.HashMap; |
| 45 | +import org.slf4j.event.KeyValuePair; |
| 46 | +import io.opentelemetry.api.common.Attributes; |
| 47 | +import io.opentelemetry.api.common.AttributeKey; |
| 48 | +import io.opentelemetry.sdk.resources.Resource; |
| 49 | +import java.util.Arrays; |
| 50 | +import java.util.List; |
| 51 | +import java.util.UUID; |
| 52 | +import java.util.Map; |
| 53 | +import org.junit.After; |
| 54 | +import org.junit.Before; |
| 55 | +import org.junit.Test; |
| 56 | + |
| 57 | +/** |
| 58 | + * Integration tests for Compute observability "golden signals". |
| 59 | + * Validates that traces, metrics, and actionable error logs are correctly recorded and exported. |
| 60 | + */ |
| 61 | +public class ITComputeGoldenSignals extends BaseTest { |
| 62 | + private static final Logger logger = (Logger) LoggerFactory.getLogger(ITComputeGoldenSignals.class); |
| 63 | + private static final String TELEMETRY_ENDPOINT = "https://telemetry.googleapis.com"; |
| 64 | + |
| 65 | + private OpenTelemetrySdk openTelemetrySdk; |
| 66 | + private TraceServiceClient traceClient; |
| 67 | + private String traceId; |
| 68 | + private String rootSpanName; |
| 69 | + private Tracer tracer; |
| 70 | + private CompositeTracerFactory compositeFactory; |
| 71 | + private InMemoryMetricReader metricReader; |
| 72 | + private TestAppender testAppender; |
| 73 | + |
| 74 | + @Before |
| 75 | + public void setUp() throws Exception { |
| 76 | + traceId = generateRandomHexString(32); |
| 77 | + rootSpanName = "ComputeRootSpan-" + generateRandomHexString(8); |
| 78 | + |
| 79 | + GoogleCredentials credentials = GoogleCredentials.getApplicationDefault(); |
| 80 | + credentials.refreshIfExpired(); |
| 81 | + String token = credentials.getAccessToken().getTokenValue(); |
| 82 | + |
| 83 | + OtlpGrpcSpanExporter spanExporter = OtlpGrpcSpanExporter.builder() |
| 84 | + .setEndpoint(TELEMETRY_ENDPOINT) |
| 85 | + .addHeader("Authorization", "Bearer " + token) |
| 86 | + .addHeader("x-goog-user-project", DEFAULT_PROJECT) |
| 87 | + .build(); |
| 88 | + |
| 89 | + BatchSpanProcessor spanProcessor = BatchSpanProcessor.builder(spanExporter).build(); |
| 90 | + |
| 91 | + Resource resource = Resource.getDefault() |
| 92 | + .merge(Resource.create(Attributes.of(AttributeKey.stringKey("gcp.project_id"), DEFAULT_PROJECT))); |
| 93 | + |
| 94 | + metricReader = InMemoryMetricReader.create(); |
| 95 | + openTelemetrySdk = OpenTelemetrySdk.builder() |
| 96 | + .setTracerProvider( |
| 97 | + SdkTracerProvider.builder() |
| 98 | + .addSpanProcessor(spanProcessor) |
| 99 | + .setResource(resource) |
| 100 | + .build()) |
| 101 | + .setMeterProvider( |
| 102 | + SdkMeterProvider.builder() |
| 103 | + .registerMetricReader(metricReader) |
| 104 | + .setResource(resource) |
| 105 | + .build()) |
| 106 | + .build(); |
| 107 | + |
| 108 | + tracer = openTelemetrySdk.getTracer("testing-compute"); |
| 109 | + |
| 110 | + // Configure TraceServiceClient with retry settings |
| 111 | + TraceServiceSettings.Builder settingsBuilder = TraceServiceSettings.newBuilder(); |
| 112 | + settingsBuilder.getTraceSettings() |
| 113 | + .setRetrySettings( |
| 114 | + RetrySettings.newBuilder() |
| 115 | + .setTotalTimeoutDuration(Duration.ofSeconds(60)) |
| 116 | + .setInitialRpcTimeoutDuration(Duration.ofSeconds(5)) |
| 117 | + .setMaxRpcTimeoutDuration(Duration.ofSeconds(10)) |
| 118 | + .build()) |
| 119 | + .setRetryableCodes(StatusCode.Code.NOT_FOUND); |
| 120 | + |
| 121 | + settingsBuilder.getStubSettingsBuilder().setTracerFactory(com.google.api.gax.tracing.BaseApiTracerFactory.getInstance()); |
| 122 | + |
| 123 | + traceClient = TraceServiceClient.create(settingsBuilder.build()); |
| 124 | + |
| 125 | + // Combine tracers using CompositeTracerFactory |
| 126 | + List<ApiTracerFactory> factories = Arrays.asList( |
| 127 | + new OpenTelemetryTracingFactory(openTelemetrySdk), |
| 128 | + new OpenTelemetryMetricsFactory(openTelemetrySdk), |
| 129 | + new LoggingTracerFactory() |
| 130 | + ); |
| 131 | + compositeFactory = new CompositeTracerFactory(factories); |
| 132 | + |
| 133 | + // Initialize and attach TestAppender |
| 134 | + testAppender = new TestAppender(); |
| 135 | + testAppender.start(); |
| 136 | + Logger loggingTracerLogger = |
| 137 | + (Logger) LoggerFactory.getLogger("com.google.api.gax.tracing.LoggingTracer"); |
| 138 | + loggingTracerLogger.addAppender(testAppender); |
| 139 | + loggingTracerLogger.setLevel(ch.qos.logback.classic.Level.DEBUG); |
| 140 | + } |
| 141 | + |
| 142 | + @After |
| 143 | + public void tearDown() throws Exception { |
| 144 | + if (traceClient != null) { |
| 145 | + traceClient.close(); |
| 146 | + } |
| 147 | + if (openTelemetrySdk != null) { |
| 148 | + openTelemetrySdk.close(); |
| 149 | + } |
| 150 | + if (testAppender != null) { |
| 151 | + ((Logger) LoggerFactory.getLogger("ROOT")) |
| 152 | + .detachAppender(testAppender); |
| 153 | + } |
| 154 | + } |
| 155 | + |
| 156 | + /** |
| 157 | + * Creates a root span with a specific trace ID to simulate an external parent context. |
| 158 | + * This helps verify that the library correctly creates child spans that inherit the parent's trace ID. |
| 159 | + * |
| 160 | + * @param traceId The trace ID to use for the root span. |
| 161 | + * @return The created root span. |
| 162 | + */ |
| 163 | + private Span createRootSpan(String traceId) { |
| 164 | + SpanContext customSpanContext = SpanContext.create(traceId, generateRandomHexString(16), TraceFlags.getSampled(), TraceState.getDefault()); |
| 165 | + return tracer.spanBuilder(rootSpanName) |
| 166 | + .setParent(Context.root().with(Span.wrap(customSpanContext))) |
| 167 | + .startSpan(); |
| 168 | + } |
| 169 | + |
| 170 | + /** |
| 171 | + * Tests that a successful compute operation generates traces that are correctly exported to Cloud Trace. |
| 172 | + */ |
| 173 | + @Test |
| 174 | + public void testComputeOperationTracing() throws Exception { |
| 175 | + String localTraceId = generateRandomHexString(32); |
| 176 | + Span rootSpan = createRootSpan(localTraceId); |
| 177 | + |
| 178 | + try (Scope scope = rootSpan.makeCurrent()) { |
| 179 | + InstancesSettings.Builder settingsBuilder = InstancesSettings.newBuilder(); |
| 180 | + settingsBuilder.getStubSettingsBuilder().setTracerFactory(compositeFactory); |
| 181 | + |
| 182 | + try (InstancesClient client = InstancesClient.create(settingsBuilder.build())) { |
| 183 | + logger.info("Listing instances in project: " + DEFAULT_PROJECT + " zone: " + DEFAULT_ZONE); |
| 184 | + client.list(DEFAULT_PROJECT, DEFAULT_ZONE); |
| 185 | + } |
| 186 | + } finally { |
| 187 | + rootSpan.end(); |
| 188 | + } |
| 189 | + |
| 190 | + openTelemetrySdk.getSdkTracerProvider().forceFlush(); |
| 191 | + fetchAndValidateTrace(localTraceId, false); |
| 192 | + validateMetrics(); |
| 193 | + validateLogging(false); |
| 194 | + } |
| 195 | + |
| 196 | + /** |
| 197 | + * Tests that a failed compute operation generates traces with error attributes. |
| 198 | + */ |
| 199 | + @Test |
| 200 | + public void testComputeOperationTracing_Error() throws Exception { |
| 201 | + String localTraceId = generateRandomHexString(32); |
| 202 | + Span rootSpan = createRootSpan(localTraceId); |
| 203 | + |
| 204 | + try (Scope scope = rootSpan.makeCurrent()) { |
| 205 | + InstancesSettings.Builder settingsBuilder = InstancesSettings.newBuilder(); |
| 206 | + settingsBuilder.getStubSettingsBuilder().setTracerFactory(compositeFactory); |
| 207 | + |
| 208 | + try (InstancesClient client = InstancesClient.create(settingsBuilder.build())) { |
| 209 | + logger.info("Triggering error by listing instances in invalid project..."); |
| 210 | + client.list("invalid-project-" + UUID.randomUUID().toString(), DEFAULT_ZONE); |
| 211 | + fail("Expected exception not thrown"); |
| 212 | + } catch (Exception e) { |
| 213 | + logger.info("Caught expected exception: " + e.getMessage()); |
| 214 | + } |
| 215 | + } finally { |
| 216 | + rootSpan.end(); |
| 217 | + } |
| 218 | + |
| 219 | + openTelemetrySdk.getSdkTracerProvider().forceFlush(); |
| 220 | + fetchAndValidateTrace(localTraceId, true); |
| 221 | + validateMetrics(); |
| 222 | + validateLogging(true); |
| 223 | + } |
| 224 | + |
| 225 | + private void fetchAndValidateTrace(String traceId, boolean expectError) throws Exception { |
| 226 | + Trace trace = traceClient.getTrace(DEFAULT_PROJECT, traceId); |
| 227 | + assertThat(trace).isNotNull(); |
| 228 | + |
| 229 | + for (TraceSpan span : trace.getSpansList()) { |
| 230 | + logger.info("Verifying attributes for span: " + span.getName()); |
| 231 | + |
| 232 | + // Skip root span as it's manually created and doesn't have RPC attributes. |
| 233 | + if (span.getName().contains("ComputeRootSpan")) { |
| 234 | + continue; |
| 235 | + } |
| 236 | + |
| 237 | + // Assert RPC span name pattern {method} {url template} |
| 238 | + assertThat(span.getName()).isEqualTo("GET compute/v1/projects/{project=*}/zones/{zone=*}/instances"); |
| 239 | + |
| 240 | + // Compute uses HTTP/REST, so we check for rpc.system.name and other HTTP attributes |
| 241 | + assertThat(span.getLabelsMap().get(ObservabilityAttributes.RPC_SYSTEM_NAME_ATTRIBUTE)).isEqualTo("http"); |
| 242 | + assertThat(span.getLabelsMap().get(ObservabilityAttributes.URL_DOMAIN_ATTRIBUTE)).isEqualTo("compute.googleapis.com"); |
| 243 | + assertThat(span.getLabelsMap().get(ObservabilityAttributes.HTTP_METHOD_ATTRIBUTE)).isEqualTo("GET"); |
| 244 | + assertThat(span.getLabelsMap().get(ObservabilityAttributes.GCP_CLIENT_SERVICE_ATTRIBUTE)).isEqualTo("compute"); |
| 245 | + assertThat(span.getLabelsMap().get(ObservabilityAttributes.REPO_ATTRIBUTE)).isEqualTo("googleapis/google-cloud-java"); |
| 246 | + |
| 247 | + if (expectError) { |
| 248 | + assertThat(span.getLabelsMap().get(ObservabilityAttributes.HTTP_RESPONSE_STATUS_ATTRIBUTE)).isEqualTo("404"); |
| 249 | + } else { |
| 250 | + assertThat(span.getLabelsMap().get(ObservabilityAttributes.HTTP_RESPONSE_STATUS_ATTRIBUTE)).isEqualTo("200"); |
| 251 | + } |
| 252 | + |
| 253 | + if (expectError) { |
| 254 | + // Verify error attributes |
| 255 | + assertThat(span.getLabelsMap()).containsKey(ObservabilityAttributes.ERROR_TYPE_ATTRIBUTE); |
| 256 | + assertThat(span.getLabelsMap()).containsKey(ObservabilityAttributes.EXCEPTION_TYPE_ATTRIBUTE); |
| 257 | + assertThat(span.getLabelsMap()).containsKey(ObservabilityAttributes.STATUS_MESSAGE_ATTRIBUTE); |
| 258 | + } |
| 259 | + } |
| 260 | + } |
| 261 | + |
| 262 | + private void validateMetrics() { |
| 263 | + Collection<MetricData> metrics = metricReader.collectAllMetrics(); |
| 264 | + logger.info("Collected " + metrics.size() + " metrics"); |
| 265 | + |
| 266 | + // GoldenSignalsMetricsRecorder.CLIENT_REQUEST_DURATION_METRIC_NAME is package-private |
| 267 | + String expectedMetricName = "gcp.client.request.duration"; |
| 268 | + |
| 269 | + MetricData durationMetric = |
| 270 | + metrics.stream() |
| 271 | + .filter(m -> m.getName().equals(expectedMetricName)) |
| 272 | + .findFirst() |
| 273 | + .orElseThrow(() -> new AssertionError("Duration metric not found: " + expectedMetricName)); |
| 274 | + |
| 275 | + logger.info("Found duration metric: " + durationMetric.getName()); |
| 276 | + |
| 277 | + // Assert that we have at least one point |
| 278 | + assertThat(durationMetric.getHistogramData().getPoints()).isNotEmpty(); |
| 279 | + |
| 280 | + io.opentelemetry.api.common.Attributes attributes = durationMetric.getHistogramData().getPoints().iterator().next().getAttributes(); |
| 281 | + assertThat(attributes.get(AttributeKey.stringKey(ObservabilityAttributes.RPC_SYSTEM_NAME_ATTRIBUTE))).isEqualTo("http"); |
| 282 | + assertThat(attributes.get(AttributeKey.stringKey(ObservabilityAttributes.GCP_CLIENT_SERVICE_ATTRIBUTE))).isEqualTo("compute"); |
| 283 | + assertThat(attributes.get(AttributeKey.stringKey(ObservabilityAttributes.URL_TEMPLATE_ATTRIBUTE))).isEqualTo("compute/v1/projects/{project=*}/zones/{zone=*}/instances"); |
| 284 | + } |
| 285 | + |
| 286 | + private void validateLogging(boolean expectError) { |
| 287 | + List<ILoggingEvent> computeEvents = new ArrayList<>(); |
| 288 | + for (ILoggingEvent event : testAppender.events) { |
| 289 | + if (event.getKeyValuePairs() == null) { |
| 290 | + continue; |
| 291 | + } |
| 292 | + Map<String, String> mdc = new HashMap<>(); |
| 293 | + for (KeyValuePair kvp : event.getKeyValuePairs()) { |
| 294 | + mdc.put(kvp.key, String.valueOf(kvp.value)); |
| 295 | + } |
| 296 | + if (!"compute".equals(mdc.get("gcp.client.service"))) { |
| 297 | + continue; |
| 298 | + } |
| 299 | + computeEvents.add(event); |
| 300 | + } |
| 301 | + |
| 302 | + if (expectError) { |
| 303 | + assertThat(computeEvents).isNotEmpty(); |
| 304 | + ILoggingEvent event = computeEvents.get(computeEvents.size() - 1); |
| 305 | + if (event.getKeyValuePairs() == null) { |
| 306 | + fail("Expected log event to have key value pairs"); |
| 307 | + } |
| 308 | + Map<String, String> mdc = new HashMap<>(); |
| 309 | + for (KeyValuePair kvp : event.getKeyValuePairs()) { |
| 310 | + mdc.put(kvp.key, String.valueOf(kvp.value)); |
| 311 | + } |
| 312 | + |
| 313 | + assertThat(mdc).containsEntry(ObservabilityAttributes.RPC_SYSTEM_NAME_ATTRIBUTE, "http"); |
| 314 | + assertThat(mdc).containsEntry(ObservabilityAttributes.GCP_CLIENT_SERVICE_ATTRIBUTE, "compute"); |
| 315 | + assertThat(mdc).containsEntry(ObservabilityAttributes.REPO_ATTRIBUTE, "googleapis/google-cloud-java"); |
| 316 | + assertThat(mdc).containsEntry(ObservabilityAttributes.HTTP_METHOD_ATTRIBUTE, "GET"); |
| 317 | + assertThat(mdc).containsKey("url.template"); |
| 318 | + assertThat(mdc).containsKey(ObservabilityAttributes.EXCEPTION_MESSAGE_ATTRIBUTE); |
| 319 | + } else { |
| 320 | + if (!computeEvents.isEmpty()) { |
| 321 | + logger.info("Captured " + computeEvents.size() + " unexpected compute log events:"); |
| 322 | + for (ILoggingEvent event : computeEvents) { |
| 323 | + logger.info("Event: " + event.getMessage() + ", Extracted: " + event.getKeyValuePairs()); |
| 324 | + } |
| 325 | + } |
| 326 | + assertThat(computeEvents).isEmpty(); |
| 327 | + } |
| 328 | + } |
| 329 | + |
| 330 | + public static class TestAppender extends AppenderBase<ILoggingEvent> { |
| 331 | + public List<ILoggingEvent> events = new ArrayList<>(); |
| 332 | + |
| 333 | + @Override |
| 334 | + protected void append(ILoggingEvent eventObject) { |
| 335 | + eventObject.getMDCPropertyMap(); |
| 336 | + events.add(eventObject); |
| 337 | + } |
| 338 | + |
| 339 | + public void clearEvents() { |
| 340 | + events.clear(); |
| 341 | + } |
| 342 | + } |
| 343 | + |
| 344 | + private String generateRandomHexString(int length) { |
| 345 | + return UUID.randomUUID().toString().replace("-", "").substring(0, length); |
| 346 | + } |
| 347 | +} |
0 commit comments