naver/pinpoint

Pinpoint is an open source APM (Application Performance Management) tool for large-scale distributed systems written in Java.

nightguide

@nightguide

pinpoint-pinpoint.log

My agent log

HyunGil Jeong

@Xylus

Hi @nightguide are you not seeing your com.cdek.pinpoint.Job's run() method at all in your trace? Are the REST-API (HTTP) , MySQL, RabbitMQ , etc from a different trace?

itymoshenko

@itymoshenko

Hello community.
I have java app that uses cassandra database to store some data interacting with db is realized using cassandra-driver-core-3.4.0 that is support in Pinpoint 1.8.5.

I have installed Pinpoint 1.8.5 in basic configuration.
And when i sending requests to app i can't see whole map and in call tree db mapped like: UNKNOWN_DB(unknown)

I have tried to specify profiler.entrypoint in agent configuration, but it's not helped.

And when i set profiler.entrypoint=com.datastax.driver.core.AbstractSession.execute, then UNKNOWN_DB(unknown) was changed to INTERNAL_METHOD in call tree.
And that's all. Destination still visualizing like unknown.

Maybe someone could point me in right direction or suggest something? Will be very appreciate for any help.

Thank you.

HyunGil Jeong

@Xylus

Hi @itymoshenko You shouldn't have to set profiler.entrypoint, the way you ran it initially is fine.

Our Cassandra integration test seems to be getting the database name okay for cassandra-driver-core-3.4.0 under pinpoint 1.8.5.

Could you tell us how you're creating cassandra connections and how you're calling cassandra?

itymoshenko

@itymoshenko

Hi @Xylus . For example, to interact with cassandra(establish connection, make some queries) we use next methods from com.datastax.driver.core:

Cluster.Builder clusterBuilder = Cluster.builder()
.addContactPoints(cassandraHosts).withPort(cassandraPort)
.withLoadBalancingPolicy(new RoundRobinPolicy())
.withPoolingOptions(getPoolingOptions());
Cluster cluster = clusterBuilder.build();

Session session = cluster.connect();

PreparedStatement pstmt = session.prepare(cql);
if(isReadQuery)
pstmt.setConsistencyLevel(readConsistencyLevel);
else
pstmt.setConsistencyLevel(writeConsistencyLevel);

BoundStatement bstmt = pstmt.bind();

ResultSet resultSet = session.execute(bstmt);
List<Row> result = resultSet.all();

itymoshenko

@itymoshenko

Hi @Xylus. Do you have any thoughts?

HyunGil Jeong

@Xylus

Hi @itymoshenko sorry for the late reply, completely missed your previous update. I'll try to replicate the issue using multiple contact points tomorrow and see if that's the issue.

On a side note, it should work fine with 1 contact point as our integration tests for Datastax driver plugin picks it up without a problem. Could you try configuring a single contact point and see if that works if possible? It'll help a lot in narrowing down the cause of the issue.

Thanks

HyunGil Jeong

@Xylus

@itymoshenko I think the problem is due to Cluster.connect(). Cassandra plugin extracts the database name from the argument passed in to Cluster.connect(String keyspace). Using Cluster.connect() without an argument, the plugin will not be able to figure out the database name.

echo

@zifeihan

hello excuse me,I use pinpoint 1.8x. the index server map has Incorrect link line about kafka plugin, un expect zero like line. like this.
through debug the pinpoint source, i notice com.navercorp.pinpoint.plugin.kafka.interceptor.ConsumerRecordEntryPointInterceptor.TraceFactoryProvider.DefaultTraceFactory#recordRootSpan(com.navercorp.pinpoint.bootstrap.context.SpanRecorder, org.apache.kafka.clients.consumer.ConsumerRecord, java.lang.String, java.lang.String) record span endPoint is kafka bootstrap.servers, i think it should record the real kafka consumer client server address.

HyunGil Jeong

@Xylus

Hi @zifeihan the issue has been resolved via github I think.

Dinesh Nithyanandam

@dinesh4747

Hi @Xylus - Need clarification on whether pinpoint starts supporting Node.js ?

I could some where it's getting called in PR or issue... Just wanted to completely confirm on tge same ?

If my understanding is right, the current version 1.8.5 doesn't have the nodejs support ? I am i right here @Xylus ? Pls clarify the same.. Thanks

Roy Kim

@RoySRose

@dinesh4747 Yes, you're right 1.8.5 doesn't have the nodejs support. To support node.js. We need node.js agent. It's being ready and will be released at the end of this year

Dinesh Nithyanandam

@dinesh4747

@RoySRose - Thanks for clarifying the same. I'll wait for the node js release in a month's time

Dinesh Nithyanandam

@dinesh4747

One more question like we have our 100 of apps sitting in GKE(Kubernetes) routing traffic via Istio, question here can we configure or ingest Pinpoint Agent as a side car or Daemonset along side application container??? @RoySRose @Xylus

Roy Kim

@RoySRose

@dinesh4747 Pinpoint will support Istio from v2.0.0.

Currently v2.0.0-RC is released. Soon the official v2.0.0. will be in few weeks.

Dinesh Nithyanandam

@dinesh4747

Thanks @RoySRose

Ankit Tomar

@anki2189

I am trying to run pinpoint-web

on port 8080

But http://localhost:8080/ is giving 404

Dinesh Nithyanandam

@dinesh4747

Does Pinpoint supports tracing of couch base client via spring boot app

Similar to what it does for Cassandra ?

yjqg6666

@yjqg6666

@RoySRose @Xylus Which version of thrift is recommended and compatible with pinpoint 1.8.5? And what about pinpoint 2.0.0-rc1?

yjqg6666

@yjqg6666

0.10?

HyunGil Jeong

@Xylus

@yjqg6666 I believe Pinpoint is using 0.10 internally. For 2.0.0, I think Pinpoint is migrating to GRPC.

@dinesh4747 Pinpoint doesn't support couch base client yet as far as I know. For Cassandra, if you're talking about Cassandra java driver, Pinpoint does have support for Cassandra Datastax java driver.

yjqg6666

@yjqg6666

ok. Thanks. from the rc1 release, thrift&grpc both are supported. would the thrift still be supported for version 2.0.0 final release?

grpc only or both?

HyunGil Jeong

@Xylus

Thrift should still be supported, but it might be going away in future versions I think

yjqg6666

@yjqg6666

ok. thanks.

yjqg6666

@yjqg6666

@Xylus One more question. If i want to make an app(in openresty + nginx+lua) to send trace data(just agent info & span data(receive&send time)) in thrift. which protocol&transportation should i use to send data to collector?

yjqg6666

@yjqg6666

@Xylus Please ignore it. I find another way to satisfy my need.

hanbum

@ParkHanbum

hello, I'm develop nodejs profiling tool. before grpc be support officially, I was try to communicate between pinpoint collector by using grpc. it was partially success, so I can record javascript function calling in nodejs to pinpoint. now I'm want try to record other metrics such as cpu usage, memory, I/O, nodejs statistics. but need some advise about protocol buffer. what and how use each field of protocol buffer? any good examples for this?

Dinesh Nithyanandam

@dinesh4747

@RoySRose - when can we expect pinpoint 2.0 release, any remote posi

Any remote possibility of early Jan 2020* ?

Jackey

@JackeyBob

@Xylus Does Pinpoint supports tracing of mongoDB？

koo-taejin

@koo-taejin

@JackeyBob Pinpoint supports mongoDB driver.
You can see support librares on this page. (https://github.com/naver/pinpoint/blob/master/doc/modules.md)

@dinesh4747 We are trying to release 2.0.0 in January.

koo-taejin

@koo-taejin

@ParkHanbum This part has a lot of data, so it might be helpful to see the code.
The codes in the com.navercorp.pinpoint.profiler.monitor.metric package are to make actual Stat messages.
If you don't understand something about a particular metric, please tell me.

Dinesh Nithyanandam

@dinesh4747

@koo-taejin - I need some real help on issue#6353 - regarding spring webflux plugin tracing...

koo-taejin

@koo-taejin

@dinesh4747 If you follow Jaehong's opinion, I expect the problem will be solved.

Dinesh Nithyanandam

@dinesh4747

@koo-taejin @Xylus @RoySRose - can someone pls take a look at issue#6353 and I have tried all the options as directed by Kim but none of them helped in getting trace for application specific method calls

Where communities thrive

People

Repo info

Activity