Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

HBASE-15242: add client side metrics for timeout and remote exceptions. #5023

Merged
merged 5 commits into from
Feb 25, 2023
Merged

HBASE-15242: add client side metrics for timeout and remote exceptions. #5023

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
b532058
HBASE-15242: add client side metrics for timeout and remote exceptions.
Feb 14, 2023
76ee7b3
Instrument all exceptions separatedly.
Feb 21, 2023
cbcc335
cosmetics changes.
Feb 21, 2023
4e74167
fix spotless.
Feb 21, 2023
b1bf32f
Fixed exception metrics with simple class name for remote exceptions.
Feb 21, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 14 additions & 2 deletions hbase-client/src/main/java/org/apache/hadoop/hbase/client/MetricsConnection.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,6 +36,8 @@
import java.util.function.Supplier;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hbase.ServerName;
import org.apache.hadoop.hbase.ipc.CallTimeoutException;
import org.apache.hadoop.hbase.ipc.RemoteWithExtrasException;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.yetus.audience.InterfaceAudience;

Expand Down Expand Up @@ -118,6 +120,9 @@ static String getScope(Configuration conf, String clusterId, Object connectionOb

private static final String CNT_BASE = "rpcCount_";
private static final String FAILURE_CNT_BASE = "rpcFailureCount_";
private static final String TIMEOUT_CNT_BASE = "rpcExceptionCallTimeout_";
private static final String REMOTE_CNT_BASE = "rpcExceptionRemote_";
vli02 marked this conversation as resolved.
Show resolved Hide resolved
private static final String OTHERS_CNT_BASE = "rpcExceptionOthers_";
vli02 marked this conversation as resolved.
Show resolved Hide resolved
private static final String DRTN_BASE = "rpcCallDurationMs_";
private static final String REQ_BASE = "rpcCallRequestSizeBytes_";
private static final String RESP_BASE = "rpcCallResponseSizeBytes_";
Expand Down Expand Up @@ -638,16 +643,23 @@ private void shutdown() {
}

/** Report RPC context to metrics system. */
public void updateRpc(MethodDescriptor method, Message param, CallStats stats, boolean failed) {
public void updateRpc(MethodDescriptor method, Message param, CallStats stats, Throwable e) {
int callsPerServer = stats.getConcurrentCallsPerServer();
if (callsPerServer > 0) {
concurrentCallsPerServerHist.update(callsPerServer);
}
// Update the counter that tracks RPCs by type.
final String methodName = method.getService().getName() + "_" + method.getName();
getMetric(CNT_BASE + methodName, rpcCounters, counterFactory).inc();
if (failed) {
if (e != null) {
getMetric(FAILURE_CNT_BASE + methodName, rpcCounters, counterFactory).inc();
if (e instanceof CallTimeoutException) {
getMetric(TIMEOUT_CNT_BASE + methodName, rpcCounters, counterFactory).inc();
} else if (e instanceof RemoteWithExtrasException) {
getMetric(REMOTE_CNT_BASE + methodName, rpcCounters, counterFactory).inc();
} else {
getMetric(OTHERS_CNT_BASE + methodName, rpcCounters, counterFactory).inc();
}
}
// this implementation is tied directly to protobuf implementation details. would be better
// if we could dispatch based on something static, ie, request Message type.
Expand Down
7 changes: 3 additions & 4 deletions hbase-client/src/main/java/org/apache/hadoop/hbase/ipc/AbstractRpcClient.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -375,16 +375,15 @@ private T getConnection(ConnectionId remoteId) throws IOException {
private void onCallFinished(Call call, HBaseRpcController hrc, Address addr,
RpcCallback<Message> callback) {
call.callStats.setCallTimeMs(EnvironmentEdgeManager.currentTime() - call.getStartTime());
final boolean failed = (call.error != null) ? true : false;
if (metrics != null) {
metrics.updateRpc(call.md, call.param, call.callStats, failed);
metrics.updateRpc(call.md, call.param, call.callStats, call.error);
}
if (LOG.isTraceEnabled()) {
LOG.trace("CallId: {}, call: {}, startTime: {}ms, callTime: {}ms, status: {}", call.id,
call.md.getName(), call.getStartTime(), call.callStats.getCallTimeMs(),
failed ? "failed" : "successful");
call.error != null ? "failed" : "successful");
}
if (failed) {
if (call.error != null) {
if (call.error instanceof RemoteException) {
call.error.fillInStackTrace();
hrc.setFailed(call.error);
Expand Down
67 changes: 58 additions & 9 deletions hbase-client/src/test/java/org/apache/hadoop/hbase/client/TestMetricsConnection.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,6 +31,8 @@
import java.util.concurrent.ThreadPoolExecutor;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hbase.HBaseClassTestRule;
import org.apache.hadoop.hbase.ipc.CallTimeoutException;
import org.apache.hadoop.hbase.ipc.RemoteWithExtrasException;
import org.apache.hadoop.hbase.security.User;
import org.apache.hadoop.hbase.testclassification.ClientTests;
import org.apache.hadoop.hbase.testclassification.MetricsTests;
Expand Down Expand Up @@ -150,51 +152,98 @@ public void testStaticMetrics() throws IOException {

for (int i = 0; i < loop; i++) {
METRICS.updateRpc(ClientService.getDescriptor().findMethodByName("Get"),
GetRequest.getDefaultInstance(), MetricsConnection.newCallStats(), false);
GetRequest.getDefaultInstance(), MetricsConnection.newCallStats(), null);
METRICS.updateRpc(ClientService.getDescriptor().findMethodByName("Scan"),
ScanRequest.getDefaultInstance(), MetricsConnection.newCallStats(), false);
ScanRequest.getDefaultInstance(), MetricsConnection.newCallStats(),
new RemoteWithExtrasException("IOException", null, false, false));
vli02 marked this conversation as resolved.
Show resolved Hide resolved
METRICS.updateRpc(ClientService.getDescriptor().findMethodByName("Multi"),
MultiRequest.getDefaultInstance(), MetricsConnection.newCallStats(), true);
MultiRequest.getDefaultInstance(), MetricsConnection.newCallStats(),
new CallTimeoutException("test with CallTimeoutException"));
METRICS.updateRpc(ClientService.getDescriptor().findMethodByName("Mutate"),
MutateRequest.newBuilder()
.setMutation(ProtobufUtil.toMutation(MutationType.APPEND, new Append(foo)))
.setRegion(region).build(),
MetricsConnection.newCallStats(), false);
MetricsConnection.newCallStats(), null);
METRICS.updateRpc(ClientService.getDescriptor().findMethodByName("Mutate"),
MutateRequest.newBuilder()
.setMutation(ProtobufUtil.toMutation(MutationType.DELETE, new Delete(foo)))
.setRegion(region).build(),
MetricsConnection.newCallStats(), false);
MetricsConnection.newCallStats(), null);
METRICS.updateRpc(ClientService.getDescriptor().findMethodByName("Mutate"),
MutateRequest.newBuilder()
.setMutation(ProtobufUtil.toMutation(MutationType.INCREMENT, new Increment(foo)))
.setRegion(region).build(),
MetricsConnection.newCallStats(), false);
MetricsConnection.newCallStats(), null);
METRICS.updateRpc(ClientService.getDescriptor().findMethodByName("Mutate"),
MutateRequest.newBuilder()
.setMutation(ProtobufUtil.toMutation(MutationType.PUT, new Put(foo))).setRegion(region)
.build(),
MetricsConnection.newCallStats(), false);
MetricsConnection.newCallStats(), null);
}

final String rpcCountPrefix = "rpcCount_" + ClientService.getDescriptor().getName() + "_";
final String rpcFailureCountPrefix =
"rpcFailureCount_" + ClientService.getDescriptor().getName() + "_";
final String rpcTimeoutCountPrefix =
"rpcExceptionCallTimeout_" + ClientService.getDescriptor().getName() + "_";
final String rpcRemoteCountPrefix =
"rpcExceptionRemote_" + ClientService.getDescriptor().getName() + "_";
String metricKey;
long metricVal;
Counter counter;
for (String method : new String[] { "Get", "Scan", "Mutate" }) {

for (String method : new String[] { "Get", "Mutate" }) {
metricKey = rpcCountPrefix + method;
metricVal = METRICS.getRpcCounters().get(metricKey).getCount();
assertTrue("metric: " + metricKey + " val: " + metricVal, metricVal >= loop);

// no failure
metricKey = rpcFailureCountPrefix + method;
counter = METRICS.getRpcCounters().get(metricKey);
metricVal = (counter != null) ? counter.getCount() : 0;
assertTrue("metric: " + metricKey + " val: " + metricVal, metricVal == 0);
}
metricKey = rpcFailureCountPrefix + "Multi";

final String scanMethod = "Scan";
metricKey = rpcCountPrefix + scanMethod;
metricVal = METRICS.getRpcCounters().get(metricKey).getCount();
assertTrue("metric: " + metricKey + " val: " + metricVal, metricVal >= loop);
// has failure
metricKey = rpcFailureCountPrefix + scanMethod;
counter = METRICS.getRpcCounters().get(metricKey);
metricVal = (counter != null) ? counter.getCount() : 0;
assertTrue("metric: " + metricKey + " val: " + metricVal, metricVal == loop);
// no timeout
metricKey = rpcTimeoutCountPrefix + scanMethod;
counter = METRICS.getRpcCounters().get(metricKey);
metricVal = (counter != null) ? counter.getCount() : 0;
assertTrue("metric: " + metricKey + " val: " + metricVal, metricVal == 0);
// has remote
metricKey = rpcRemoteCountPrefix + scanMethod;
counter = METRICS.getRpcCounters().get(metricKey);
metricVal = (counter != null) ? counter.getCount() : 0;
assertTrue("metric: " + metricKey + " val: " + metricVal, metricVal == loop);

final String multiMethod = "Multi";
metricKey = rpcCountPrefix + multiMethod;
metricVal = METRICS.getRpcCounters().get(metricKey).getCount();
assertTrue("metric: " + metricKey + " val: " + metricVal, metricVal >= loop);
// has failure
metricKey = rpcFailureCountPrefix + multiMethod;
counter = METRICS.getRpcCounters().get(metricKey);
metricVal = (counter != null) ? counter.getCount() : 0;
assertTrue("metric: " + metricKey + " val: " + metricVal, metricVal == loop);
// has timeout
metricKey = rpcTimeoutCountPrefix + multiMethod;
counter = METRICS.getRpcCounters().get(metricKey);
metricVal = (counter != null) ? counter.getCount() : 0;
assertTrue("metric: " + metricKey + " val: " + metricVal, metricVal == loop);
// no remote
metricKey = rpcRemoteCountPrefix + multiMethod;
counter = METRICS.getRpcCounters().get(metricKey);
metricVal = (counter != null) ? counter.getCount() : 0;
assertTrue("metric: " + metricKey + " val: " + metricVal, metricVal == 0);

for (MetricsConnection.CallTracker t : new MetricsConnection.CallTracker[] {
METRICS.getGetTracker(), METRICS.getScanTracker(), METRICS.getMultiTracker(),
METRICS.getAppendTracker(), METRICS.getDeleteTracker(), METRICS.getIncrementTracker(),
Expand Down