Search in sources :

Example 16 with ComponentMetrics

use of com.microsoft.dhalion.metrics.ComponentMetrics in project incubator-heron by apache.

the class BufferSizeSensorTest method registerStMgrInstanceMetricResponse.

static void registerStMgrInstanceMetricResponse(MetricsProvider metricsProvider, String metric, long value) {
    Map<String, ComponentMetrics> result = new HashMap<>();
    ComponentMetrics metrics = new ComponentMetrics("__stmgr__");
    InstanceMetrics instanceMetrics = new InstanceMetrics("stmgr-1");
    instanceMetrics.addMetric(metric, value);
    metrics.addInstanceMetric(instanceMetrics);
    result.put("__stmgr__", metrics);
    when(metricsProvider.getComponentMetrics(metric, DEFAULT_METRIC_DURATION, "__stmgr__")).thenReturn(result);
}
Also used : InstanceMetrics(com.microsoft.dhalion.metrics.InstanceMetrics) HashMap(java.util.HashMap) ComponentMetrics(com.microsoft.dhalion.metrics.ComponentMetrics)

Example 17 with ComponentMetrics

use of com.microsoft.dhalion.metrics.ComponentMetrics in project incubator-heron by apache.

the class MetricsCacheMetricsProviderTest method testGetTimeLineMetrics.

@Test
public void testGetTimeLineMetrics() {
    MetricsCacheMetricsProvider spyMetricsProvider = createMetricsProviderSpy();
    String metric = "count";
    String comp = "bolt";
    TopologyMaster.MetricResponse response = TopologyMaster.MetricResponse.newBuilder().setStatus(Status.newBuilder().setStatus(StatusCode.OK)).addMetric(TaskMetric.newBuilder().setInstanceId("container_1_bolt_1").addMetric(IndividualMetric.newBuilder().setName(metric).addIntervalValues(IntervalValue.newBuilder().setValue("104").setInterval(MetricInterval.newBuilder().setStart(1497481288).setEnd(1497481288))))).addMetric(TaskMetric.newBuilder().setInstanceId("container_1_bolt_2").addMetric(IndividualMetric.newBuilder().setName(metric).addIntervalValues(IntervalValue.newBuilder().setValue("12").setInterval(MetricInterval.newBuilder().setStart(1497481228).setEnd(1497481228))).addIntervalValues(IntervalValue.newBuilder().setValue("2").setInterval(MetricInterval.newBuilder().setStart(1497481348).setEnd(1497481348))).addIntervalValues(IntervalValue.newBuilder().setValue("3").setInterval(MetricInterval.newBuilder().setStart(1497481168).setEnd(1497481168))))).build();
    doReturn(response).when(spyMetricsProvider).getMetricsFromMetricsCache(metric, comp, Instant.ofEpochSecond(10), Duration.ofSeconds(60));
    Map<String, ComponentMetrics> metrics = spyMetricsProvider.getComponentMetrics(metric, Instant.ofEpochSecond(10), Duration.ofSeconds(60), comp);
    assertEquals(1, metrics.size());
    ComponentMetrics componentMetrics = metrics.get(comp);
    assertNotNull(componentMetrics);
    assertEquals(2, componentMetrics.getMetrics().size());
    InstanceMetrics instanceMetrics = componentMetrics.getMetrics("container_1_bolt_1");
    assertNotNull(instanceMetrics);
    assertEquals(1, instanceMetrics.getMetrics().size());
    Map<Instant, Double> metricValues = instanceMetrics.getMetrics().get(metric);
    assertEquals(1, metricValues.size());
    assertEquals(104, metricValues.get(Instant.ofEpochSecond(1497481288)).intValue());
    instanceMetrics = componentMetrics.getMetrics("container_1_bolt_2");
    assertNotNull(instanceMetrics);
    assertEquals(1, instanceMetrics.getMetrics().size());
    metricValues = instanceMetrics.getMetrics().get(metric);
    assertEquals(3, metricValues.size());
    assertEquals(12, metricValues.get(Instant.ofEpochSecond(1497481228L)).intValue());
    assertEquals(2, metricValues.get(Instant.ofEpochSecond(1497481348L)).intValue());
    assertEquals(3, metricValues.get(Instant.ofEpochSecond(1497481168L)).intValue());
}
Also used : InstanceMetrics(com.microsoft.dhalion.metrics.InstanceMetrics) Instant(java.time.Instant) TopologyMaster(com.twitter.heron.proto.tmaster.TopologyMaster) ComponentMetrics(com.microsoft.dhalion.metrics.ComponentMetrics) Test(org.junit.Test)

Example 18 with ComponentMetrics

use of com.microsoft.dhalion.metrics.ComponentMetrics in project incubator-heron by apache.

the class BackPressureDetector method detect.

/**
 * Detects all components initiating backpressure above the configured limit. Normally there
 * will be only one component
 *
 * @return A collection of all components causing backpressure.
 */
@Override
public List<Symptom> detect() {
    ArrayList<Symptom> result = new ArrayList<>();
    Map<String, ComponentMetrics> backpressureMetrics = bpSensor.get();
    for (ComponentMetrics compMetrics : backpressureMetrics.values()) {
        ComponentMetricsHelper compStats = new ComponentMetricsHelper(compMetrics);
        compStats.computeBpStats();
        if (compStats.getTotalBackpressure() > noiseFilterMillis) {
            LOG.info(String.format("Detected back pressure for %s, total back pressure is %f", compMetrics.getName(), compStats.getTotalBackpressure()));
            result.add(new Symptom(SYMPTOM_BACK_PRESSURE.text(), compMetrics));
        }
    }
    return result;
}
Also used : ArrayList(java.util.ArrayList) ComponentMetricsHelper(com.twitter.heron.healthmgr.common.ComponentMetricsHelper) Symptom(com.microsoft.dhalion.detector.Symptom) ComponentMetrics(com.microsoft.dhalion.metrics.ComponentMetrics)

Example 19 with ComponentMetrics

use of com.microsoft.dhalion.metrics.ComponentMetrics in project incubator-heron by apache.

the class DataSkewDiagnoser method diagnose.

@Override
public Diagnosis diagnose(List<Symptom> symptoms) {
    List<Symptom> bpSymptoms = getBackPressureSymptoms(symptoms);
    Map<String, ComponentMetrics> processingRateSkewComponents = getProcessingRateSkewComponents(symptoms);
    Map<String, ComponentMetrics> waitQDisparityComponents = getWaitQDisparityComponents(symptoms);
    if (bpSymptoms.isEmpty() || processingRateSkewComponents.isEmpty() || waitQDisparityComponents.isEmpty()) {
        // Since there is no back pressure or disparate execute count, no action is needed
        return null;
    } else if (bpSymptoms.size() > 1) {
        // TODO handle cases where multiple detectors create back pressure symptom
        throw new IllegalStateException("Multiple back-pressure symptoms case");
    }
    ComponentMetrics bpMetrics = bpSymptoms.iterator().next().getComponent();
    // verify data skew, larger queue size and back pressure for the same component exists
    ComponentMetrics exeCountMetrics = processingRateSkewComponents.get(bpMetrics.getName());
    ComponentMetrics pendingBufferMetrics = waitQDisparityComponents.get(bpMetrics.getName());
    if (exeCountMetrics == null || pendingBufferMetrics == null) {
        // for the component with back pressure. This is not a data skew case
        return null;
    }
    ComponentMetrics mergedData = ComponentMetrics.merge(bpMetrics, ComponentMetrics.merge(exeCountMetrics, pendingBufferMetrics));
    ComponentMetricsHelper compStats = new ComponentMetricsHelper(mergedData);
    compStats.computeBpStats();
    MetricsStats exeStats = compStats.computeMinMaxStats(METRIC_EXE_COUNT);
    MetricsStats bufferStats = compStats.computeMinMaxStats(METRIC_BUFFER_SIZE);
    Symptom resultSymptom = null;
    for (InstanceMetrics boltMetrics : compStats.getBoltsWithBackpressure()) {
        double exeCount = boltMetrics.getMetricValueSum(METRIC_EXE_COUNT.text());
        double bufferSize = boltMetrics.getMetricValueSum(METRIC_BUFFER_SIZE.text());
        double bpValue = boltMetrics.getMetricValueSum(METRIC_BACK_PRESSURE.text());
        if (exeStats.getMetricMax() < 1.10 * exeCount && bufferStats.getMetricMax() < 2 * bufferSize) {
            LOG.info(String.format("DataSkew: %s back-pressure(%s), high execution count: %s and " + "high buffer size %s", boltMetrics.getName(), bpValue, exeCount, bufferSize));
            resultSymptom = new Symptom(SYMPTOM_DATA_SKEW.text(), mergedData);
        }
    }
    return resultSymptom != null ? new Diagnosis(DIAGNOSIS_DATA_SKEW.text(), resultSymptom) : null;
}
Also used : InstanceMetrics(com.microsoft.dhalion.metrics.InstanceMetrics) ComponentMetricsHelper(com.twitter.heron.healthmgr.common.ComponentMetricsHelper) Diagnosis(com.microsoft.dhalion.diagnoser.Diagnosis) Symptom(com.microsoft.dhalion.detector.Symptom) ComponentMetrics(com.microsoft.dhalion.metrics.ComponentMetrics) MetricsStats(com.twitter.heron.healthmgr.common.MetricsStats)

Example 20 with ComponentMetrics

use of com.microsoft.dhalion.metrics.ComponentMetrics in project incubator-heron by apache.

the class UnderProvisioningDiagnoser method diagnose.

@Override
public Diagnosis diagnose(List<Symptom> symptoms) {
    List<Symptom> bpSymptoms = getBackPressureSymptoms(symptoms);
    Map<String, ComponentMetrics> processingRateSkewComponents = getProcessingRateSkewComponents(symptoms);
    Map<String, ComponentMetrics> waitQDisparityComponents = getWaitQDisparityComponents(symptoms);
    if (bpSymptoms.isEmpty() || !processingRateSkewComponents.isEmpty() || !waitQDisparityComponents.isEmpty()) {
        // and buffer sizes, no action is needed
        return null;
    } else if (bpSymptoms.size() > 1) {
        // TODO handle cases where multiple detectors create back pressure symptom
        throw new IllegalStateException("Multiple back-pressure symptoms case");
    }
    ComponentMetrics bpMetrics = bpSymptoms.iterator().next().getComponent();
    ComponentMetricsHelper compStats = new ComponentMetricsHelper(bpMetrics);
    compStats.computeBpStats();
    LOG.info(String.format("UNDER_PROVISIONING: %s back-pressure(%s) and similar processing rates " + "and buffer sizes", bpMetrics.getName(), compStats.getTotalBackpressure()));
    Symptom resultSymptom = new Symptom(SYMPTOM_UNDER_PROVISIONING.text(), bpMetrics);
    return new Diagnosis(DIAGNOSIS_UNDER_PROVISIONING.text(), resultSymptom);
}
Also used : ComponentMetricsHelper(com.twitter.heron.healthmgr.common.ComponentMetricsHelper) Diagnosis(com.microsoft.dhalion.diagnoser.Diagnosis) Symptom(com.microsoft.dhalion.detector.Symptom) ComponentMetrics(com.microsoft.dhalion.metrics.ComponentMetrics)

Aggregations

ComponentMetrics (com.microsoft.dhalion.metrics.ComponentMetrics)29 Symptom (com.microsoft.dhalion.detector.Symptom)16 InstanceMetrics (com.microsoft.dhalion.metrics.InstanceMetrics)15 Test (org.junit.Test)15 HashMap (java.util.HashMap)14 ComponentMetricsHelper (com.twitter.heron.healthmgr.common.ComponentMetricsHelper)7 Diagnosis (com.microsoft.dhalion.diagnoser.Diagnosis)6 HealthPolicyConfig (com.twitter.heron.healthmgr.HealthPolicyConfig)6 ArrayList (java.util.ArrayList)6 MetricsStats (com.twitter.heron.healthmgr.common.MetricsStats)4 Instant (java.time.Instant)4 MetricsProvider (com.microsoft.dhalion.api.MetricsProvider)3 PackingPlanProvider (com.twitter.heron.healthmgr.common.PackingPlanProvider)3 TopologyProvider (com.twitter.heron.healthmgr.common.TopologyProvider)3 BufferSizeSensor (com.twitter.heron.healthmgr.sensors.BufferSizeSensor)3 Action (com.microsoft.dhalion.resolver.Action)2 ExecuteCountSensor (com.twitter.heron.healthmgr.sensors.ExecuteCountSensor)2 TopologyMaster (com.twitter.heron.proto.tmaster.TopologyMaster)2 PackingPlan (com.twitter.heron.spi.packing.PackingPlan)2 TopologyAPI (com.twitter.heron.api.generated.TopologyAPI)1