Concurrency HTTP sessions for `OtlpHttpClient` #1209

owent · 2022-02-13T11:05:09Z

Fixes #1176

Changes

Allow multiple running http sessions in OtlpHttpClient
Add concurrency_sessions to control the max concurrency http sessions.

For significant contributions please make sure you have completed the following items:

CHANGELOG.md updated for non-trivial changes
Unit tests have been added
Changes in public API reviewed

codecov · 2022-02-13T11:11:59Z

Codecov Report

Merging #1209 (075bc27) into main (31d888a) will decrease coverage by 0.30%.
The diff coverage is 80.72%.

@@            Coverage Diff             @@
##             main    #1209      +/-   ##
==========================================
- Coverage   91.99%   91.69%   -0.29%     
==========================================
  Files         205      205              
  Lines        7395     7528     +133     
==========================================
+ Hits         6802     6902     +100     
- Misses        593      626      +33

Impacted Files	Coverage Δ
...lemetry/exporters/memory/in_memory_span_exporter.h	`75.00% <0.00%> (-16.30%)`	⬇️
...de/opentelemetry/exporters/ostream/span_exporter.h	`100.00% <ø> (ø)`
exporters/ostream/src/span_exporter.cc	`87.26% <0.00%> (-3.56%)`	⬇️
...nclude/opentelemetry/ext/http/client/http_client.h	`93.34% <ø> (ø)`
sdk/include/opentelemetry/sdk/trace/exporter.h	`100.00% <ø> (ø)`
sdk/test/trace/simple_processor_test.cc	`77.42% <0.00%> (-8.29%)`	⬇️
api/include/opentelemetry/common/timestamp.h	`81.25% <25.00%> (-18.75%)`	⬇️
...ntelemetry/ext/http/client/curl/http_client_curl.h	`87.81% <35.72%> (-4.50%)`	⬇️
...include/opentelemetry/sdk/trace/simple_processor.h	`86.96% <57.15%> (-13.04%)`	⬇️
sdk/src/trace/batch_span_processor.cc	`93.44% <94.74%> (+0.41%)`	⬆️
... and 3 more

lalitb · 2022-02-17T07:14:10Z

Thanks for the PR. Would be reviewing this early next week.

owent · 2022-02-19T13:04:52Z

Thanks for the PR. Would be reviewing this early next week.

After #1185 is merged, there are no unit test for concurrency sessions now. I wonder whether should I add unit test for this, if so, the connecting problem (in CI / Bazel valgrind) may happen again.

esigo · 2022-02-19T20:26:15Z

ext/include/opentelemetry/ext/http/client/curl/http_client_curl.h

-      session.second->CancelSession();
+      std::map<uint64_t, std::shared_ptr<Session>> sessions;
+      sessions.swap(sessions_);
+      for (auto &session : sessions)


why iterating over sessions_ is not ok but iterating over sessions is ok here

The call stack HttpClient::CancelAllSessions -> Session::CancelSession -> HttpClient::CleanupSession will crash here.
CleanupSession will change sessions_ when iterating it.

esigo · 2022-02-19T20:30:00Z

exporters/otlp/test/otlp_http_log_exporter_test.cc

@@ -256,6 +259,8 @@ TEST_F(OtlpHttpLogExporterTestPeer, ExportBinaryIntegrationTest)
        http_client::nosend::Response response;
        callback.OnResponse(response);


don't we need response.Finish(callback); here?

I want to reuse it to trigger OnEvent and call OnResponse. The codes do not trigger OnEvent before.

exporters/otlp/include/opentelemetry/exporters/otlp/otlp_http_client.h

esigo · 2022-02-19T21:00:35Z

exporters/otlp/src/otlp_http_client.cc

-  request->SetMethod(http_client::Method::Post);
-  request->SetBody(body_vec);
-  request->ReplaceHeader("Content-Type", content_type);
+    std::lock_guard<std::recursive_mutex> guard{session_manager_lock_};


Why do we need to do all the processing before checking for Shutdown?

It's used to protect isShutdown(), is_shutdown_, running_sessions_ and http_client_. They are all not thread-safe.

I think we should make isShutdown thread safe, and short circuit here (e.g. return if isShutdown() is true).

esigo · 2022-02-19T21:11:34Z

exporters/otlp/src/otlp_http_client.cc

 bool OtlpHttpClient::isShutdown() const noexcept
 {
-  const std::lock_guard<opentelemetry::common::SpinLockMutex> locked(lock_);
  return is_shutdown_;


Without lock this read would have race condition.

is_shutdown_ and isShutdown() are protected by session_manager_lock_ now.

owent · 2022-02-20T12:04:26Z

The timeout test seems not a problem of this PR

(//sdk/test/common:circular_buffer_benchmark_smoketest under valgrind)

owent · 2022-02-23T07:11:50Z

We have a internal benmark report which shows this PR increase QPS of OtlpHttpLogExporter from 4k/second to 13k/second.

lalitb · 2022-02-23T09:31:11Z

exporters/otlp/src/otlp_http_client.cc

+  std::unique_lock<std::mutex> lock(session_waker_lock_);
+  bool wait_successful = session_waker_.wait_for(lock, options_.timeout, [this] {
+    std::lock_guard<std::recursive_mutex> guard{session_manager_lock_};
+    return running_sessions_.size() <= options_.concurrent_sessions;


What will happen if number of current running sessions are less than max-configured? Export will return success even though there is no real successful export ?

Yes.
We have a service need to export a lot logs.When the QPS is greater than 4k, the thread of otel-cpp cost only 7% of CPU time but it start to drop datas. Because when the exporter is waiting for http response, the batch log processor is still receiving more logs and the queue in it grow quickly and will be full soon.
In our test,after we set concurrent session to 8, it can send more than 13k request/second and drop nothing.

Yes, I can see the improvement we can achieve with these changes. The concern I have is that this deviates from the specs (https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/trace/sdk.md#interface-definition-1), which states that the Export method should return the result of transfer of the data over the wire. This can allow the Span Processor to take decision accordingly say retrying the failed transactions.
We may want to change the Export method to return the result callback to make these changes compliant to specs ?

Also, had a quick look into how other SIGs are handling this. Java returns export status to processor asynchronously using CompletableFuture, while JS returns status using callback.

FYI https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/protocol/otlp.md#otlpgrpc-concurrent-requests

Also, had a quick look into how other SIGs are handling this. Java returns export status to processor asynchronously using CompletableFuture, while JS returns status using callback.

Good idea. Maybe we can use nostd::variant<std::future<sdk::common::ExportResult>> here. But it's a break change for exporters's APIs. I think it need more discussion about it.

FYI https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/protocol/otlp.md#otlpgrpc-concurrent-requests

The OtlpGrpcExporter and OtlpGrpcLogExporter have the same problem, but I didn't try to use asynchronize API or test the performance yet.

But it's a break change for exporters's APIs. I think it need more discussion about it.

Yes, I was thinking about that too. We have to keep supporting the existing Exporter::Export() interface and add new using callback/future. And a new config option for SpanProcessor to select which interface to use. But it needs some thought and discussion.

The OtlpGrpcExporter and OtlpGrpcLogExporter have the same problem, but I didn't try to use asynchronize API or test the performance yet.

Agree, they too suffer from the same issue.

lalitb · 2022-02-23T09:32:43Z

We have a internal benmark report which shows this PR increase QPS of OtlpHttpLogExporter from 4k/second to 13k/second.

Good to know that. I had quick glance at the Export() function, will review it more thoroughly tomorrow. Sorry for the delay on this.

api/include/opentelemetry/common/timestamp.h

reyang · 2022-02-23T17:35:07Z

exporters/otlp/include/opentelemetry/exporters/otlp/otlp_http_client.h

+   *
+   * @return return true if there are more sessions to delete
+   */
+  bool cleanupGCSessions() noexcept;


What is "GC"?

GC here mean garbage collection of http sessions.When a http session is finished, we can not destroy it immediately right now, because it may still be visited in some codes. (e.g. When we remove a session in OnResponse, it will be visited in OnEvent later).
So I move the finished sessions into gc_sessions_ and real destroy it soon later.

ThomsonTan · 2022-02-23T18:15:18Z

exporters/otlp/include/opentelemetry/exporters/otlp/otlp_http_client.h

@@ -99,6 +109,8 @@ class OtlpHttpClient
   */
  explicit OtlpHttpClient(OtlpHttpClientOptions &&options);

+  ~OtlpHttpClient();


Is this needed to be explicit?

Yes, we need to hold the lock and then clean up all sessions. When removing sessions, we must make sure not to change gc_sessions_ when we iterating it to call FinishSession of all running sessions.

lalitb · 2022-02-23T22:29:44Z

@owent - This is a great initiative, thanks once again for the PR. Do you think it would make sense to have exporter return export status as callback and let the processor manage the concurrent export magic? This will ensure we are specs compliant, and also not much changes would be required to support Zipkin exporter. And later, we can make it truly asynchronous by adding support of multi_handle for curl client.

+ Temporary fix the `Export()` may be called too many times when we shutdown or just wakeup worker thread once.This problem is completely fixed in open-telemetry#1209 but not merged yet. + Fix http client may has a different error code when our network is under a proxy.

…ction (open-telemetry#1278)

Signed-off-by: owent <admin@owent.net>

…pHttpLogExporter`. Add tests for both sync and async exporting. Signed-off-by: owent <admin@owent.net>

Signed-off-by: owent <admin@owent.net>

…alled without any span and log in async mode Signed-off-by: owent <admin@owent.net>

…ecords when we shudown batch log/span processor. Signed-off-by: owent <admin@owent.net>

Signed-off-by: owent <admin@owent.net>

lalitb · 2022-03-23T17:19:02Z

@owent - can we close this PR if the changes are already in feature branch?

lalitb · 2022-03-23T17:19:02Z

@owent - can we close this PR if the changes are already in feature branch?

owent · 2022-03-24T02:12:11Z

@owent - can we close this PR if the changes are already in feature branch?

Yes, this is already moved to async-changes

lalitb · 2022-05-03T18:42:41Z

@owent fyi - open-telemetry/opentelemetry-specification#2452 (comment)

owent requested a review from a team February 13, 2022 11:05

owent changed the title ~~\[WIP\] Cocurrency http sessions for OtlpHttpClient~~ [WIP] Cocurrency http sessions for OtlpHttpClient Feb 13, 2022

owent changed the title ~~[WIP] Cocurrency http sessions for OtlpHttpClient~~ Cocurrency http sessions for OtlpHttpClient Feb 14, 2022

owent force-pushed the cocurrency_otlp_http_session branch from 1cafb98 to 073c598 Compare February 15, 2022 03:06

owent mentioned this pull request Feb 19, 2022

Unit test failing in CI: OtlpHttpExporterTestPeer.ExportJsonIntegrationTest #1049

Closed

owent force-pushed the cocurrency_otlp_http_session branch from 000162c to db327de Compare February 19, 2022 12:55

esigo reviewed Feb 19, 2022

View reviewed changes

exporters/otlp/include/opentelemetry/exporters/otlp/otlp_http_client.h Outdated Show resolved Hide resolved

esigo reviewed Feb 19, 2022

View reviewed changes

lalitb reviewed Feb 23, 2022

View reviewed changes

reyang changed the title ~~Cocurrency http sessions for OtlpHttpClient~~ Concurrency HTTP sessions for OtlpHttpClient Feb 23, 2022

reyang reviewed Feb 23, 2022

View reviewed changes

api/include/opentelemetry/common/timestamp.h Outdated Show resolved Hide resolved

reyang reviewed Feb 23, 2022

View reviewed changes

api/include/opentelemetry/common/timestamp.h Outdated Show resolved Hide resolved

reyang reviewed Feb 23, 2022

View reviewed changes

ThomsonTan reviewed Feb 23, 2022

View reviewed changes

owent force-pushed the cocurrency_otlp_http_session branch from 6c8838e to 6ae64e5 Compare February 24, 2022 12:44

esigo mentioned this pull request Mar 1, 2022

Consider the performance harm of synchronized exporters #1238

Closed

owent force-pushed the cocurrency_otlp_http_session branch from 6ae64e5 to fca159b Compare March 4, 2022 03:13

owent mentioned this pull request Mar 4, 2022

Using libcurl-multi to implement opentelemetry::ext::http::client::curl::HttpClient #1243

Closed

owent mentioned this pull request Mar 8, 2022

Reorder the destructor of members in LoggerProvider and TracerProvider #1245

Merged

3 tasks

owent mentioned this pull request Mar 8, 2022

Async/callback based interface between Span/Log Processor and Span/Log Exporter #1239

Closed

owent force-pushed the cocurrency_otlp_http_session branch 2 times, most recently from 659e11f to e47da76 Compare March 13, 2022 05:31

owent mentioned this pull request Mar 13, 2022

Added changes for async callback mechanism from Processor to Exporter #1252

Closed

3 tasks

owent force-pushed the cocurrency_otlp_http_session branch from e47da76 to 90942be Compare March 21, 2022 05:40

owent mentioned this pull request Mar 21, 2022

Cocurrency otlp http session #1274

Merged

3 tasks

owent and others added 2 commits March 20, 2022 23:41

Cocurrency otlp http session (open-telemetry#1274)

c3eaa9d

Async callback Exporter to Processor changes (open-telemetry#1275)

6f53da3

owent force-pushed the cocurrency_otlp_http_session branch from 90942be to 521831f Compare March 22, 2022 08:12

DebajitDas and others added 2 commits March 22, 2022 11:17

Changing the type of callback function in Export function to std::fun…

729c2f8

…ction (open-telemetry#1278)

Implement async version of OtlpHttpClient::Export

4c3dc25

Signed-off-by: owent <admin@owent.net>

owent force-pushed the cocurrency_otlp_http_session branch from 521831f to 4174407 Compare March 23, 2022 04:32

Finish async Export for OtlpHttpClient, OtlpHttpExporter and `Otl…

8ed3424

…pHttpLogExporter`. Add tests for both sync and async exporting. Signed-off-by: owent <admin@owent.net>

owent force-pushed the cocurrency_otlp_http_session branch from 4174407 to 8ed3424 Compare March 23, 2022 04:35

owent added 4 commits March 23, 2022 15:36

Fix always waiting when ForceFlush is called when shutdown.

5c2598e

Signed-off-by: owent <admin@owent.net>

Fix missing calling NotifyForceFlushCompletion when ForceFlush is c…

815e4e0

…alled without any span and log in async mode Signed-off-by: owent <admin@owent.net>

Fix missing calling to NotifyShutdownCompletion, when there is no r…

8659418

…ecords when we shudown batch log/span processor. Signed-off-by: owent <admin@owent.net>

Fix the lifetime of resource referenced by async exporting callback

075bc27

Signed-off-by: owent <admin@owent.net>

owent closed this Mar 24, 2022

lalitb mentioned this pull request Mar 24, 2022

Batch span processor support for multiple pending exports open-telemetry/opentelemetry-specification#2434

Closed

owent mentioned this pull request May 20, 2022

Merge all commits of async-changes into main #1413

Merged

3 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Concurrency HTTP sessions for `OtlpHttpClient` #1209

Concurrency HTTP sessions for `OtlpHttpClient` #1209

owent commented Feb 13, 2022 •

edited

Loading

codecov bot commented Feb 13, 2022 •

edited

Loading

lalitb commented Feb 17, 2022

owent commented Feb 19, 2022

esigo Feb 19, 2022

owent Feb 20, 2022

esigo Feb 19, 2022

owent Feb 20, 2022

esigo Feb 19, 2022 •

edited

Loading

owent Feb 20, 2022

reyang Feb 23, 2022

esigo Feb 19, 2022

owent Feb 20, 2022

owent commented Feb 20, 2022 •

edited

Loading

owent commented Feb 23, 2022

lalitb Feb 23, 2022

owent Feb 23, 2022 •

edited

Loading

lalitb Feb 23, 2022 •

edited

Loading

lalitb Feb 23, 2022

reyang Feb 23, 2022

owent Feb 24, 2022 •

edited

Loading

owent Feb 24, 2022

lalitb Feb 24, 2022 •

edited

Loading

lalitb Feb 24, 2022

lalitb commented Feb 23, 2022

reyang Feb 23, 2022

owent Feb 24, 2022

ThomsonTan Feb 23, 2022

owent Feb 24, 2022 •

edited

Loading

lalitb commented Feb 23, 2022 •

edited

Loading

lalitb commented Mar 23, 2022

lalitb commented Mar 23, 2022

owent commented Mar 24, 2022

lalitb commented May 3, 2022

		@@ -256,6 +259,8 @@ TEST_F(OtlpHttpLogExporterTestPeer, ExportBinaryIntegrationTest)
		http_client::nosend::Response response;
		callback.OnResponse(response);

Concurrency HTTP sessions for OtlpHttpClient #1209

Concurrency HTTP sessions for OtlpHttpClient #1209

Conversation

owent commented Feb 13, 2022 • edited Loading

Changes

codecov bot commented Feb 13, 2022 • edited Loading

Codecov Report

lalitb commented Feb 17, 2022

owent commented Feb 19, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

esigo Feb 19, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

owent commented Feb 20, 2022 • edited Loading

owent commented Feb 23, 2022

Choose a reason for hiding this comment

owent Feb 23, 2022 • edited Loading

Choose a reason for hiding this comment

lalitb Feb 23, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

owent Feb 24, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lalitb Feb 24, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lalitb commented Feb 23, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

owent Feb 24, 2022 • edited Loading

Choose a reason for hiding this comment

lalitb commented Feb 23, 2022 • edited Loading

lalitb commented Mar 23, 2022

lalitb commented Mar 23, 2022

owent commented Mar 24, 2022

lalitb commented May 3, 2022

Concurrency HTTP sessions for `OtlpHttpClient` #1209

Concurrency HTTP sessions for `OtlpHttpClient` #1209

owent commented Feb 13, 2022 •

edited

Loading

codecov bot commented Feb 13, 2022 •

edited

Loading

esigo Feb 19, 2022 •

edited

Loading

owent commented Feb 20, 2022 •

edited

Loading

owent Feb 23, 2022 •

edited

Loading

lalitb Feb 23, 2022 •

edited

Loading

owent Feb 24, 2022 •

edited

Loading

lalitb Feb 24, 2022 •

edited

Loading

owent Feb 24, 2022 •

edited

Loading

lalitb commented Feb 23, 2022 •

edited

Loading