scwhittle commented on code in PR #30211:
URL: https://github.com/apache/beam/pull/30211#discussion_r1480458218


##########
runners/google-cloud-dataflow-java/worker/src/main/java/org/apache/beam/runners/dataflow/worker/status/ChannelzServlet.java:
##########
@@ -0,0 +1,245 @@
+package org.apache.beam.runners.dataflow.worker.status;
+
+import org.apache.beam.vendor.grpc.v1p60p1.io.grpc.channelz.v1.*;
+import 
org.apache.beam.vendor.grpc.v1p60p1.io.grpc.protobuf.services.ChannelzService;
+import org.apache.beam.vendor.grpc.v1p60p1.io.grpc.stub.StreamObserver;
+import 
org.apache.beam.vendor.guava.v32_1_2_jre.com.google.common.util.concurrent.SettableFuture;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import javax.servlet.ServletException;
+import javax.servlet.http.HttpServletRequest;
+import javax.servlet.http.HttpServletResponse;
+import java.io.IOException;
+import java.io.PrintWriter;
+import java.util.HashSet;
+import java.util.List;
+import java.util.Set;
+
+/** Respond to /channelz with the GRPC channelz data. */
+@SuppressWarnings({
+        "nullness" // TODO(https://github.com/apache/beam/issues/20497)
+})
+public class ChannelzServlet extends BaseStatusServlet implements 
DebugCapture.Capturable {
+    private static final String PATH = "/channelz";
+    private static final Logger LOG = 
LoggerFactory.getLogger(ChannelzServlet.class);
+    private static final int MAX_TOP_CHANNELS_TO_RETURN = 100;
+
+    private final ChannelzService channelzService = 
ChannelzService.newInstance(MAX_TOP_CHANNELS_TO_RETURN);
+
+    public ChannelzServlet() {
+        super(PATH);
+    }
+
+    @Override
+    protected void doGet(HttpServletRequest request, HttpServletResponse 
response) throws IOException, ServletException {
+        response.setStatus(HttpServletResponse.SC_OK);
+        PrintWriter writer = response.getWriter();
+        captureData(writer);
+    }
+
+    @Override
+    public String pageName() {
+        return PATH;
+    }
+
+    @Override
+    public void captureData(PrintWriter writer) {
+        writer.println("<html>");
+        writer.println("<h1>Channelz</h1>");
+        appendTopChannels(writer);
+        writer.println("</html>");
+    }
+
+    // channelz proto says there may not be cycles in the ref graph

Review Comment:
   nit: period at end of comment
   also a little confusing, does it say there may be cycles? or are we just 
being extra careful to prevent a possible cycle?



##########
runners/google-cloud-dataflow-java/worker/src/main/java/org/apache/beam/runners/dataflow/worker/status/ChannelzServlet.java:
##########
@@ -0,0 +1,245 @@
+package org.apache.beam.runners.dataflow.worker.status;
+
+import org.apache.beam.vendor.grpc.v1p60p1.io.grpc.channelz.v1.*;
+import 
org.apache.beam.vendor.grpc.v1p60p1.io.grpc.protobuf.services.ChannelzService;
+import org.apache.beam.vendor.grpc.v1p60p1.io.grpc.stub.StreamObserver;
+import 
org.apache.beam.vendor.guava.v32_1_2_jre.com.google.common.util.concurrent.SettableFuture;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import javax.servlet.ServletException;
+import javax.servlet.http.HttpServletRequest;
+import javax.servlet.http.HttpServletResponse;
+import java.io.IOException;
+import java.io.PrintWriter;
+import java.util.HashSet;
+import java.util.List;
+import java.util.Set;
+
+/** Respond to /channelz with the GRPC channelz data. */
+@SuppressWarnings({
+        "nullness" // TODO(https://github.com/apache/beam/issues/20497)

Review Comment:
   can this be fixed up to not require this?



##########
runners/google-cloud-dataflow-java/worker/src/main/java/org/apache/beam/runners/dataflow/worker/status/ChannelzServlet.java:
##########
@@ -0,0 +1,245 @@
+package org.apache.beam.runners.dataflow.worker.status;
+
+import org.apache.beam.vendor.grpc.v1p60p1.io.grpc.channelz.v1.*;
+import 
org.apache.beam.vendor.grpc.v1p60p1.io.grpc.protobuf.services.ChannelzService;
+import org.apache.beam.vendor.grpc.v1p60p1.io.grpc.stub.StreamObserver;
+import 
org.apache.beam.vendor.guava.v32_1_2_jre.com.google.common.util.concurrent.SettableFuture;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import javax.servlet.ServletException;
+import javax.servlet.http.HttpServletRequest;
+import javax.servlet.http.HttpServletResponse;
+import java.io.IOException;
+import java.io.PrintWriter;
+import java.util.HashSet;
+import java.util.List;
+import java.util.Set;
+
+/** Respond to /channelz with the GRPC channelz data. */
+@SuppressWarnings({
+        "nullness" // TODO(https://github.com/apache/beam/issues/20497)
+})
+public class ChannelzServlet extends BaseStatusServlet implements 
DebugCapture.Capturable {
+    private static final String PATH = "/channelz";
+    private static final Logger LOG = 
LoggerFactory.getLogger(ChannelzServlet.class);
+    private static final int MAX_TOP_CHANNELS_TO_RETURN = 100;
+
+    private final ChannelzService channelzService = 
ChannelzService.newInstance(MAX_TOP_CHANNELS_TO_RETURN);
+
+    public ChannelzServlet() {
+        super(PATH);
+    }
+
+    @Override
+    protected void doGet(HttpServletRequest request, HttpServletResponse 
response) throws IOException, ServletException {
+        response.setStatus(HttpServletResponse.SC_OK);
+        PrintWriter writer = response.getWriter();
+        captureData(writer);
+    }
+
+    @Override
+    public String pageName() {
+        return PATH;
+    }
+
+    @Override
+    public void captureData(PrintWriter writer) {
+        writer.println("<html>");
+        writer.println("<h1>Channelz</h1>");
+        appendTopChannels(writer);
+        writer.println("</html>");
+    }
+
+    // channelz proto says there may not be cycles in the ref graph
+    // we track visited ids to prevent any accidental cycles,
+    static class VisitedSets {
+        Set<Long> channels = new HashSet<>();
+        Set<Long> subchannels = new HashSet<>();
+    }
+
+    private void appendTopChannels(PrintWriter writer) {
+        SettableFuture<GetTopChannelsResponse> future = 
SettableFuture.create();
+        channelzService.getTopChannels(
+                GetTopChannelsRequest.newBuilder()

Review Comment:
   how about filtering the channels to the dataflow streaming endpoints?
   
   If we want to allow printing others to aid with debugging client library 
stuckness we could perhaps have an option modifying the filter.  But by default 
let's just show the dataflow streaming ones.



##########
runners/google-cloud-dataflow-java/worker/src/main/java/org/apache/beam/runners/dataflow/worker/status/ChannelzServlet.java:
##########
@@ -0,0 +1,245 @@
+package org.apache.beam.runners.dataflow.worker.status;
+
+import org.apache.beam.vendor.grpc.v1p60p1.io.grpc.channelz.v1.*;
+import 
org.apache.beam.vendor.grpc.v1p60p1.io.grpc.protobuf.services.ChannelzService;
+import org.apache.beam.vendor.grpc.v1p60p1.io.grpc.stub.StreamObserver;
+import 
org.apache.beam.vendor.guava.v32_1_2_jre.com.google.common.util.concurrent.SettableFuture;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import javax.servlet.ServletException;
+import javax.servlet.http.HttpServletRequest;
+import javax.servlet.http.HttpServletResponse;
+import java.io.IOException;
+import java.io.PrintWriter;
+import java.util.HashSet;
+import java.util.List;
+import java.util.Set;
+
+/** Respond to /channelz with the GRPC channelz data. */
+@SuppressWarnings({
+        "nullness" // TODO(https://github.com/apache/beam/issues/20497)
+})
+public class ChannelzServlet extends BaseStatusServlet implements 
DebugCapture.Capturable {
+    private static final String PATH = "/channelz";
+    private static final Logger LOG = 
LoggerFactory.getLogger(ChannelzServlet.class);
+    private static final int MAX_TOP_CHANNELS_TO_RETURN = 100;
+
+    private final ChannelzService channelzService = 
ChannelzService.newInstance(MAX_TOP_CHANNELS_TO_RETURN);
+
+    public ChannelzServlet() {
+        super(PATH);
+    }
+
+    @Override
+    protected void doGet(HttpServletRequest request, HttpServletResponse 
response) throws IOException, ServletException {
+        response.setStatus(HttpServletResponse.SC_OK);
+        PrintWriter writer = response.getWriter();
+        captureData(writer);
+    }
+
+    @Override
+    public String pageName() {
+        return PATH;
+    }
+
+    @Override
+    public void captureData(PrintWriter writer) {
+        writer.println("<html>");
+        writer.println("<h1>Channelz</h1>");
+        appendTopChannels(writer);
+        writer.println("</html>");
+    }
+
+    // channelz proto says there may not be cycles in the ref graph
+    // we track visited ids to prevent any accidental cycles,
+    static class VisitedSets {
+        Set<Long> channels = new HashSet<>();
+        Set<Long> subchannels = new HashSet<>();
+    }
+
+    private void appendTopChannels(PrintWriter writer) {
+        SettableFuture<GetTopChannelsResponse> future = 
SettableFuture.create();
+        channelzService.getTopChannels(
+                GetTopChannelsRequest.newBuilder()
+                        .build(), getStreamObserver(future));
+        GetTopChannelsResponse topChannelsResponse;
+        try {
+            topChannelsResponse = future.get();
+        } catch (Exception e) {
+            String msg = "Failed to get channelz: " + e.getMessage();
+            LOG.warn(msg, e);
+            writer.println(msg);
+            return;
+        }
+        writer.println("<h2>Top Level Channels</h2>");
+        writer.println("<table border='1'>");
+        VisitedSets visitedSets = new VisitedSets();
+        for (Channel channel : topChannelsResponse.getChannelList()) {
+            writer.println("<tr>");
+            writer.println("<td>");
+            writer.println("TopChannelId: " + channel.getRef().getChannelId());
+            writer.println("</td>");
+            writer.println("<td>");
+            appendChannel(channel, writer, visitedSets);
+            writer.println("</td>");
+            writer.println("</tr>");
+        }
+        writer.println("</table>");
+    }
+
+
+
+    private void appendChannels(List<ChannelRef> channelRefs, PrintWriter 
writer, VisitedSets visitedSets) {
+        for (ChannelRef channelRef : channelRefs) {
+            writer.println("<tr>");
+            writer.println("<td>");
+            writer.println("Channel: " + channelRef.getChannelId());
+            writer.println("</td>");
+            writer.println("<td>");
+            appendChannel(channelRef, writer, visitedSets);
+            writer.println("</td>");
+            writer.println("</tr>");
+        }
+    }
+
+    private void appendChannel(ChannelRef channelRef, PrintWriter writer, 
VisitedSets visitedSets) {
+        if (visitedSets.channels.contains(channelRef.getChannelId())) {
+            String msg = "Duplicate Channel Id: " + channelRef;
+            LOG.warn(msg);
+            writer.println(msg);
+            return;
+        }
+        visitedSets.channels.add(channelRef.getChannelId());
+        SettableFuture<GetChannelResponse> future = SettableFuture.create();
+        channelzService.getChannel(
+                
GetChannelRequest.newBuilder().setChannelId(channelRef.getChannelId()).build(),
+                getStreamObserver(future));
+        Channel channel;
+        try {
+            channel = future.get().getChannel();
+        } catch (Exception e) {
+            String msg = "Failed to get Channel: " + channelRef;
+            LOG.warn(msg, e);
+            writer.println(msg + " Exception: " + e.getMessage());
+            return;
+        }
+        appendChannel(channel, writer, visitedSets);
+    }
+
+    private void appendChannel(Channel channel, PrintWriter writer, 
VisitedSets visitedSets) {
+        writer.println("<table border='1'>");
+        writer.println("<tr>");
+        writer.println("<td>");
+        writer.println("ChannelId: " + channel.getRef().getChannelId());
+        writer.println("</td>");
+        writer.println("<td><pre>" + channel);
+        writer.println("</pre></td>");
+        writer.println("</tr>");
+        appendChannels(channel.getChannelRefList(), writer, visitedSets);
+        appendSubChannels(channel.getSubchannelRefList(), writer, visitedSets);
+        appendSockets(channel.getSocketRefList(), writer);
+        writer.println("</table>");
+    }
+
+    private void appendSubChannels(List<SubchannelRef> subchannelRefList, 
PrintWriter writer, VisitedSets visitedSets) {
+        for (SubchannelRef subchannelRef : subchannelRefList) {
+            writer.println("<tr>");
+            writer.println("<td>");
+            writer.println("Sub Channel: " + subchannelRef.getSubchannelId());
+            writer.println("</td>");
+            writer.println("<td>");
+            appendSubchannel(subchannelRef, writer, visitedSets);
+            writer.println("</td>");
+            writer.println("</tr>");
+        }
+    }
+
+    private void appendSubchannel(SubchannelRef subchannelRef, PrintWriter 
writer, VisitedSets visitedSets) {
+        if (visitedSets.subchannels.contains(subchannelRef.getSubchannelId())) 
{
+            String msg = "Duplicate Subchannel Id: " + subchannelRef;
+            LOG.warn(msg);
+            writer.println(msg);
+            return;
+        }
+        visitedSets.subchannels.add(subchannelRef.getSubchannelId());
+        SettableFuture<GetSubchannelResponse> future = SettableFuture.create();
+        channelzService.getSubchannel(
+                
GetSubchannelRequest.newBuilder().setSubchannelId(subchannelRef.getSubchannelId())
+                        .build(),
+                getStreamObserver(future));
+        Subchannel subchannel;
+        try {
+            subchannel = future.get().getSubchannel();
+        } catch (Exception e) {
+            String msg = "Failed to get Subchannel: " + subchannelRef;
+            LOG.warn(msg, e);

Review Comment:
   not sure it's worth logging, can just show it in the output page
   
   ditto for others



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to