YT-21938: Introduce get_job_trace

* Changelog entry\ Type: feature\ Component: proxy Add the handler that gets the job’s trace events. commit_hash:c9d1a6d57e886a266967d57065bd3948882808c7
author: omgronny <omgronny@yandex-team.com> 2024-10-11 21:16:54 +0300
committer: omgronny <omgronny@yandex-team.com> 2024-10-11 21:28:33 +0300
commit: c7307fc40d61e2a996fd7c22ee9ef52663897b3b (patch)
tree: cf38637a1ef654e984c4847260253dd8195bff66 /yt
parent: 1df197e6035ea9826bfedee7d48812e318ba9c7a (diff)
download: ydb-c7307fc40d61e2a996fd7c22ee9ef52663897b3b.tar.gz
26 files changed, 382 insertions, 1 deletions
diff --git a/yt/cpp/mapreduce/client/client.cpp b/yt/cpp/mapreduce/client/client.cpp
index 9e2885bbe0..9dab176bde 100644
--- a/yt/cpp/mapreduce/client/client.cpp
+++ b/yt/cpp/mapreduce/client/client.cpp
@@ -1243,6 +1243,14 @@ IFileReaderPtr TClient::GetJobStderr(
     return NRawClient::GetJobStderr(Context_, operationId, jobId, options);
 }
 
+std::vector<TJobTraceEvent> TClient::GetJobTrace(
+    const TOperationId& operationId,
+    const TGetJobTraceOptions& options)
+{
+    CheckShutdown();
+    return NRawClient::GetJobTrace(ClientRetryPolicy_->CreatePolicyForGenericRequest(), Context_, operationId, options);
+}
+
 TNode::TListType TClient::SkyShareTable(
     const std::vector<TYPath>& tablePaths,
     const TSkyShareTableOptions& options)
diff --git a/yt/cpp/mapreduce/client/client.h b/yt/cpp/mapreduce/client/client.h
index 5de00285ef..32c458316d 100644
--- a/yt/cpp/mapreduce/client/client.h
+++ b/yt/cpp/mapreduce/client/client.h
@@ -449,6 +449,10 @@ public:
         const TJobId& jobId,
         const TGetJobStderrOptions& options = TGetJobStderrOptions()) override;
 
+    std::vector<TJobTraceEvent> GetJobTrace(
+        const TOperationId& operationId,
+        const TGetJobTraceOptions& options = TGetJobTraceOptions()) override;
+
     TNode::TListType SkyShareTable(
         const std::vector<TYPath>& tablePaths,
         const TSkyShareTableOptions& options = TSkyShareTableOptions()) override;
diff --git a/yt/cpp/mapreduce/interface/client.h b/yt/cpp/mapreduce/interface/client.h
index 56efa3c23c..3032025140 100644
--- a/yt/cpp/mapreduce/interface/client.h
+++ b/yt/cpp/mapreduce/interface/client.h
@@ -493,6 +493,18 @@ public:
         const TGetJobStderrOptions& options = TGetJobStderrOptions()) = 0;
 
     ///
+    /// @brief Get trace of a job.
+    ///
+    /// @ref NYT::TErrorResponse exception is thrown if it is missing.
+    ///
+    /// @note YT doesn't store all job traces.
+    ///
+    /// @see [YT doc](https://ytsaurus.tech/docs/en/api/commands.html#get_job_trace)
+    virtual std::vector<TJobTraceEvent> GetJobTrace(
+        const TOperationId& operationId,
+        const TGetJobTraceOptions& options = TGetJobTraceOptions()) = 0;
+
+    ///
     /// @brief Create one or several rbtorrents for files in a blob table.
     ///
     /// If specified, one torrent is created for each value of `KeyColumns` option.
diff --git a/yt/cpp/mapreduce/interface/fwd.h b/yt/cpp/mapreduce/interface/fwd.h
index 0434c03d8b..485b45129a 100644
--- a/yt/cpp/mapreduce/interface/fwd.h
+++ b/yt/cpp/mapreduce/interface/fwd.h
@@ -157,6 +157,7 @@ namespace NYT {
     using TTabletCellId = TGUID;
     using TReplicaId = TGUID;
     using TJobId = TGUID;
+    using TJobTraceId = TGUID;
 
     using TYPath = TString;
     using TLocalFilePath = TString;
@@ -370,6 +371,8 @@ namespace NYT {
 
     struct TListJobsOptions;
 
+    struct TGetJobTraceOptions;
+
     struct IOperationClient;
 
     enum class EFinishedJobState : int;
diff --git a/yt/cpp/mapreduce/interface/operation.h b/yt/cpp/mapreduce/interface/operation.h
index 9a85049886..f2de3ea3bd 100644
--- a/yt/cpp/mapreduce/interface/operation.h
+++ b/yt/cpp/mapreduce/interface/operation.h
@@ -3048,6 +3048,70 @@ struct TGetFailedJobInfoOptions
 ////////////////////////////////////////////////////////////////////////////////
 
 ///
+/// @brief Options for @ref NYT::IClient::GetJobTrace.
+struct TGetJobTraceOptions
+{
+    /// @cond Doxygen_Suppress
+    using TSelf = TGetJobTraceOptions;
+    /// @endcond
+
+    ///
+    /// @brief Id of the job.
+    FLUENT_FIELD_OPTION(TJobId, JobId);
+
+    ///
+    /// @brief Id of the trace.
+    FLUENT_FIELD_OPTION(TJobTraceId, TraceId);
+
+    ///
+    /// @brief Search for traces with time >= `FromTime`.
+    FLUENT_FIELD_OPTION(i64, FromTime);
+
+    ///
+    /// @brief Search for traces with time <= `ToTime`.
+    FLUENT_FIELD_OPTION(i64, ToTime);
+
+    ///
+    /// @brief Search for traces with event index >= `FromEventIndex`.
+    FLUENT_FIELD_OPTION(i64, FromEventIndex);
+
+    ///
+    /// @brief Search for traces with event index >= `ToEventIndex`.
+    FLUENT_FIELD_OPTION(i64, ToEventIndex);
+};
+
+///
+/// @brief Response for @ref NYT::IOperation::GetJobTrace.
+struct TJobTraceEvent
+{
+    ///
+    /// @brief Id of the operation.
+    TOperationId OperationId;
+
+    ///
+    /// @brief Id of the job.
+    TJobId JobId;
+
+    ///
+    /// @brief Id of the trace.
+    TJobTraceId TraceId;
+
+    ///
+    /// @brief Index of the trace event.
+    i64 EventIndex;
+
+    ///
+    /// @brief Raw evenr in json format.
+    TString Event;
+
+    ///
+    /// @brief Time of the event.
+    TInstant EventTime;
+};
+
+////////////////////////////////////////////////////////////////////////////////
+
+///
 /// @brief Interface representing an operation.
 struct IOperation
     : public TThrRefBase
diff --git a/yt/cpp/mapreduce/raw_client/raw_requests.cpp b/yt/cpp/mapreduce/raw_client/raw_requests.cpp
index 2f9610a1ca..59868c599e 100644
--- a/yt/cpp/mapreduce/raw_client/raw_requests.cpp
+++ b/yt/cpp/mapreduce/raw_client/raw_requests.cpp
@@ -783,6 +783,55 @@ IFileReaderPtr GetJobStderr(
     return new TResponseReader(context, std::move(header));
 }
 
+TJobTraceEvent ParseJobTraceEvent(const TNode& node)
+{
+    const auto& mapNode = node.AsMap();
+    TJobTraceEvent result;
+
+    if (auto idNode = mapNode.FindPtr("operation_id")) {
+        result.OperationId = GetGuid(idNode->AsString());
+    }
+    if (auto idNode = mapNode.FindPtr("job_id")) {
+        result.JobId = GetGuid(idNode->AsString());
+    }
+    if (auto idNode = mapNode.FindPtr("trace_id")) {
+        result.TraceId = GetGuid(idNode->AsString());
+    }
+    if (auto eventIndexNode = mapNode.FindPtr("event_index")) {
+        result.EventIndex = eventIndexNode->AsInt64();
+    }
+    if (auto eventNode = mapNode.FindPtr("event")) {
+        result.Event = eventNode->AsString();
+    }
+    if (auto eventTimeNode = mapNode.FindPtr("event_time")) {
+        result.EventTime = TInstant::ParseIso8601(eventTimeNode->AsString());;
+    }
+
+    return result;
+}
+
+std::vector<TJobTraceEvent> GetJobTrace(
+    const IRequestRetryPolicyPtr& retryPolicy,
+    const TClientContext& context,
+    const TOperationId& operationId,
+    const TGetJobTraceOptions& options)
+{
+    THttpHeader header("GET", "get_job_trace");
+    header.MergeParameters(SerializeParamsForGetJobTrace(operationId, options));
+    auto responseInfo = RetryRequestWithPolicy(retryPolicy, context, header);
+    auto resultNode = NodeFromYsonString(responseInfo.Response);
+
+    std::vector<TJobTraceEvent> result;
+
+    const auto& traceEventNodesList = resultNode.AsList();
+    result.reserve(traceEventNodesList.size());
+    for (const auto& traceEventNode : traceEventNodesList) {
+        result.push_back(ParseJobTraceEvent(traceEventNode));
+    }
+
+    return result;
+}
+
 TMaybe<TYPath> GetFileFromCache(
     const IRequestRetryPolicyPtr& retryPolicy,
     const TClientContext& context,
diff --git a/yt/cpp/mapreduce/raw_client/raw_requests.h b/yt/cpp/mapreduce/raw_client/raw_requests.h
index c2d1a53b51..0a183d617c 100644
--- a/yt/cpp/mapreduce/raw_client/raw_requests.h
+++ b/yt/cpp/mapreduce/raw_client/raw_requests.h
@@ -257,6 +257,12 @@ TString GetJobStderrWithRetries(
     const TJobId& jobId,
     const TGetJobStderrOptions& options = TGetJobStderrOptions());
 
+std::vector<TJobTraceEvent> GetJobTrace(
+    const IRequestRetryPolicyPtr& retryPolicy,
+    const TClientContext& context,
+    const TOperationId& operationId,
+    const TGetJobTraceOptions& options = TGetJobTraceOptions());
+
 //
 // File cache
 //
diff --git a/yt/cpp/mapreduce/raw_client/rpc_parameters_serialization.cpp b/yt/cpp/mapreduce/raw_client/rpc_parameters_serialization.cpp
index a638ac581d..bebc59ec91 100644
--- a/yt/cpp/mapreduce/raw_client/rpc_parameters_serialization.cpp
+++ b/yt/cpp/mapreduce/raw_client/rpc_parameters_serialization.cpp
@@ -525,6 +525,15 @@ TNode SerializeParamsForGetJob(
     return result;
 }
 
+TNode SerializeParamsForGetJobTrace(
+    const TOperationId& operationId,
+    const TGetJobTraceOptions& /* options */)
+{
+    TNode result;
+    SetOperationIdParam(&result, operationId);
+    return result;
+}
+
 TNode SerializeParamsForListJobs(
     const TOperationId& operationId,
     const TListJobsOptions& options)
diff --git a/yt/cpp/mapreduce/raw_client/rpc_parameters_serialization.h b/yt/cpp/mapreduce/raw_client/rpc_parameters_serialization.h
index a7ab35d91d..69a4888267 100644
--- a/yt/cpp/mapreduce/raw_client/rpc_parameters_serialization.h
+++ b/yt/cpp/mapreduce/raw_client/rpc_parameters_serialization.h
@@ -126,6 +126,10 @@ TNode SerializeParamsForListJobs(
     const TOperationId& operationId,
     const TListJobsOptions& options);
 
+TNode SerializeParamsForGetJobTrace(
+    const TOperationId& operationId,
+    const TGetJobTraceOptions& options);
+
 TNode SerializeParametersForInsertRows(
     const TString& pathPrefix,
     const TYPath& path,
diff --git a/yt/yt/client/api/delegating_client.h b/yt/yt/client/api/delegating_client.h
index c14f2d49d6..5b5e9accf3 100644
--- a/yt/yt/client/api/delegating_client.h
+++ b/yt/yt/client/api/delegating_client.h
@@ -516,6 +516,11 @@ public:
         const TGetJobStderrOptions& options),
         (operationIdOrAlias, jobId, options))
 
+    DELEGATE_METHOD(TFuture<std::vector<TJobTraceEvent>>, GetJobTrace, (
+        const NScheduler::TOperationIdOrAlias& operationIdOrAlias,
+        const TGetJobTraceOptions& options),
+        (operationIdOrAlias, options))
+
     DELEGATE_METHOD(TFuture<TSharedRef>, GetJobFailContext, (
         const NScheduler::TOperationIdOrAlias& operationIdOrAlias,
         NJobTrackerClient::TJobId jobId,
diff --git a/yt/yt/client/api/operation_client.cpp b/yt/yt/client/api/operation_client.cpp
index 4e1816a6e7..8f8dbb355b 100644
--- a/yt/yt/client/api/operation_client.cpp
+++ b/yt/yt/client/api/operation_client.cpp
@@ -217,6 +217,19 @@ void Serialize(const TJob& job, NYson::IYsonConsumer* consumer, TStringBuf idKey
         .EndMap();
 }
 
+void Serialize(const TJobTraceEvent& traceEvent, NYson::IYsonConsumer* consumer)
+{
+    NYTree::BuildYsonFluently(consumer)
+        .BeginMap()
+            .Item("operation_id").Value(traceEvent.OperationId)
+            .Item("job_id").Value(traceEvent.JobId)
+            .Item("trace_id").Value(traceEvent.TraceId)
+            .Item("event_index").Value(traceEvent.EventIndex)
+            .Item("event").Value(traceEvent.Event)
+            .Item("event_time").Value(traceEvent.EventTime.MicroSeconds())
+        .EndMap();
+}
+
 ////////////////////////////////////////////////////////////////////////////////
 
 void TListOperationsAccessFilter::Register(TRegistrar registrar)
diff --git a/yt/yt/client/api/operation_client.h b/yt/yt/client/api/operation_client.h
index 48687497da..d5072db1fc 100644
--- a/yt/yt/client/api/operation_client.h
+++ b/yt/yt/client/api/operation_client.h
@@ -89,6 +89,18 @@ struct TGetJobStderrOptions
     std::optional<i64> Offset;
 };
 
+struct TGetJobTraceOptions
+    : public TTimeoutOptions
+    , public TMasterReadOptions
+{
+    std::optional<NJobTrackerClient::TJobId> JobId;
+    std::optional<NScheduler::TJobTraceId> TraceId;
+    std::optional<i64> FromTime;
+    std::optional<i64> ToTime;
+    std::optional<i64> FromEventIndex;
+    std::optional<i64> ToEventIndex;
+};
+
 struct TGetJobFailContextOptions
     : public TTimeoutOptions
     , public TMasterReadOptions
@@ -346,6 +358,18 @@ struct TJob
 
 void Serialize(const TJob& job, NYson::IYsonConsumer* consumer, TStringBuf idKey);
 
+struct TJobTraceEvent
+{
+    NJobTrackerClient::TOperationId OperationId;
+    NJobTrackerClient::TJobId JobId;
+    NScheduler::TJobTraceId TraceId;
+    i64 EventIndex;
+    TString Event;
+    TInstant EventTime;
+};
+
+void Serialize(const TJobTraceEvent& traceEvent, NYson::IYsonConsumer* consumer);
+
 struct TListJobsStatistics
 {
     TEnumIndexedArray<NJobTrackerClient::EJobState, i64> StateCounts;
@@ -443,6 +467,10 @@ struct IOperationClient
         NJobTrackerClient::TJobId jobId,
         const TGetJobStderrOptions& options = {}) = 0;
 
+    virtual TFuture<std::vector<TJobTraceEvent>> GetJobTrace(
+        const NScheduler::TOperationIdOrAlias& operationIdOrAlias,
+        const TGetJobTraceOptions& options = {}) = 0;
+
     virtual TFuture<TSharedRef> GetJobFailContext(
         const NScheduler::TOperationIdOrAlias& operationIdOrAlias,
         NJobTrackerClient::TJobId jobId,
diff --git a/yt/yt/client/api/rpc_proxy/api_service_proxy.h b/yt/yt/client/api/rpc_proxy/api_service_proxy.h
index 6b00c7470a..4dcb04c33f 100644
--- a/yt/yt/client/api/rpc_proxy/api_service_proxy.h
+++ b/yt/yt/client/api/rpc_proxy/api_service_proxy.h
@@ -121,6 +121,7 @@ public:
     DEFINE_RPC_PROXY_METHOD(NRpcProxy::NProto, GetJobInputPaths);
     DEFINE_RPC_PROXY_METHOD(NRpcProxy::NProto, GetJobSpec);
     DEFINE_RPC_PROXY_METHOD(NRpcProxy::NProto, GetJobStderr);
+    DEFINE_RPC_PROXY_METHOD(NRpcProxy::NProto, GetJobTrace);
     DEFINE_RPC_PROXY_METHOD(NRpcProxy::NProto, GetJobFailContext);
     DEFINE_RPC_PROXY_METHOD(NRpcProxy::NProto, AbandonJob);
     DEFINE_RPC_PROXY_METHOD(NRpcProxy::NProto, PollJobShell);
diff --git a/yt/yt/client/api/rpc_proxy/client_impl.cpp b/yt/yt/client/api/rpc_proxy/client_impl.cpp
index b9e4944381..50bf51f247 100644
--- a/yt/yt/client/api/rpc_proxy/client_impl.cpp
+++ b/yt/yt/client/api/rpc_proxy/client_impl.cpp
@@ -1328,6 +1328,41 @@ TFuture<TGetJobStderrResponse> TClient::GetJobStderr(
     }));
 }
 
+TFuture<std::vector<TJobTraceEvent>> TClient::GetJobTrace(
+    const TOperationIdOrAlias& operationIdOrAlias,
+    const TGetJobTraceOptions& options)
+{
+    auto proxy = CreateApiServiceProxy();
+
+    auto req = proxy.GetJobTrace();
+    SetTimeoutOptions(*req, options);
+
+    NScheduler::ToProto(req, operationIdOrAlias);
+    if (options.JobId) {
+        ToProto(req->mutable_job_id(), *options.JobId);
+    }
+    if (options.TraceId) {
+        ToProto(req->mutable_trace_id(), *options.TraceId);
+    }
+    if (options.FromTime) {
+        req->set_from_time(*options.FromTime);
+    }
+    if (options.ToTime) {
+        req->set_to_time(*options.ToTime);
+    }
+    if (options.FromEventIndex) {
+        req->set_from_event_index(*options.FromEventIndex);
+    }
+    if (options.ToEventIndex) {
+        req->set_to_event_index(*options.ToEventIndex);
+    }
+
+    return req->Invoke().Apply(BIND([] (const TApiServiceProxy::TRspGetJobTracePtr& rsp) {
+        return FromProto<std::vector<TJobTraceEvent>>(rsp->events());
+    }));
+}
+
+
 TFuture<TSharedRef> TClient::GetJobFailContext(
     const TOperationIdOrAlias& operationIdOrAlias,
     NJobTrackerClient::TJobId jobId,
diff --git a/yt/yt/client/api/rpc_proxy/client_impl.h b/yt/yt/client/api/rpc_proxy/client_impl.h
index d3905300e3..9f8ee540a2 100644
--- a/yt/yt/client/api/rpc_proxy/client_impl.h
+++ b/yt/yt/client/api/rpc_proxy/client_impl.h
@@ -278,6 +278,10 @@ public:
         NJobTrackerClient::TJobId jobId,
         const NApi::TGetJobStderrOptions& options) override;
 
+    TFuture<std::vector<TJobTraceEvent>> GetJobTrace(
+        const NScheduler::TOperationIdOrAlias& operationIdOrAlias,
+        const NApi::TGetJobTraceOptions& options) override;
+
     TFuture<TSharedRef> GetJobFailContext(
         const NScheduler::TOperationIdOrAlias& operationIdOrAlias,
         NJobTrackerClient::TJobId jobId,
diff --git a/yt/yt/client/api/rpc_proxy/helpers.cpp b/yt/yt/client/api/rpc_proxy/helpers.cpp
index a89368573c..af3b4b7eb6 100644
--- a/yt/yt/client/api/rpc_proxy/helpers.cpp
+++ b/yt/yt/client/api/rpc_proxy/helpers.cpp
@@ -426,6 +426,30 @@ void FromProto(
     FromProto(&result->Errors, proto.errors());
 }
 
+void ToProto(
+    NProto::TJobTraceEvent* proto,
+    const NApi::TJobTraceEvent& result)
+{
+    ToProto(proto->mutable_operation_id(), result.OperationId);
+    ToProto(proto->mutable_job_id(), result.JobId);
+    ToProto(proto->mutable_trace_id(), result.TraceId);
+    proto->set_event_index(result.EventIndex);
+    proto->set_event(result.Event);
+    proto->set_event_time(ToProto<i64>(result.EventTime));
+}
+
+void FromProto(
+    NApi::TJobTraceEvent* result,
+    const NProto::TJobTraceEvent& proto)
+{
+    FromProto(&result->OperationId, proto.operation_id());
+    FromProto(&result->JobId, proto.job_id());
+    FromProto(&result->TraceId, proto.trace_id());
+    result->EventIndex = proto.event_index();
+    result->Event = proto.event();
+    result->EventTime = TInstant::FromValue(proto.event_time());
+}
+
 ////////////////////////////////////////////////////////////////////////////////
 // MISC
 ////////////////////////////////////////////////////////////////////////////////
diff --git a/yt/yt/client/api/rpc_proxy/helpers.h b/yt/yt/client/api/rpc_proxy/helpers.h
index 9027337359..ca99986c72 100644
--- a/yt/yt/client/api/rpc_proxy/helpers.h
+++ b/yt/yt/client/api/rpc_proxy/helpers.h
@@ -106,6 +106,14 @@ void FromProto(
     NApi::TListJobsResult* result,
     const NProto::TListJobsResult& proto);
 
+void ToProto(
+    NProto::TJobTraceEvent* proto,
+    const NApi::TJobTraceEvent& result);
+
+void FromProto(
+    NApi::TJobTraceEvent* result,
+    const NProto::TJobTraceEvent& proto);
+
 void ToProto(NProto::TColumnSchema* protoSchema, const NTableClient::TColumnSchema& schema);
 void FromProto(NTableClient::TColumnSchema* schema, const NProto::TColumnSchema& protoSchema);
 
diff --git a/yt/yt/client/driver/driver.cpp b/yt/yt/client/driver/driver.cpp
index 19a0f3bdc5..4fb7783697 100644
--- a/yt/yt/client/driver/driver.cpp
+++ b/yt/yt/client/driver/driver.cpp
@@ -292,6 +292,7 @@ public:
         REGISTER_ALL(TGetJobInputCommand,                  "get_job_input",                   Null,       Binary,     false, true );
         REGISTER_ALL(TGetJobInputPathsCommand,             "get_job_input_paths",             Null,       Structured, false, true );
         REGISTER_ALL(TGetJobStderrCommand,                 "get_job_stderr",                  Null,       Binary,     false, true );
+        REGISTER_ALL(TGetJobTraceCommand,                  "get_job_trace",                   Null,       Structured, false, true );
         REGISTER_ALL(TGetJobFailContextCommand,            "get_job_fail_context",            Null,       Binary,     false, true );
         REGISTER_ALL(TGetJobSpecCommand,                   "get_job_spec",                    Null,       Structured, false, true );
         REGISTER_ALL(TListOperationsCommand,               "list_operations",                 Null,       Structured, false, false);
diff --git a/yt/yt/client/driver/scheduler_commands.cpp b/yt/yt/client/driver/scheduler_commands.cpp
index 9cb7396725..53f82c7bf2 100644
--- a/yt/yt/client/driver/scheduler_commands.cpp
+++ b/yt/yt/client/driver/scheduler_commands.cpp
@@ -174,6 +174,52 @@ void TGetJobStderrCommand::DoExecute(ICommandContextPtr context)
 
 ////////////////////////////////////////////////////////////////////////////////
 
+void TGetJobTraceCommand::Register(TRegistrar registrar)
+{
+    registrar.ParameterWithUniversalAccessor<std::optional<TJobId>>(
+        "job_id",
+        [] (TThis* command) -> auto& {return command->Options.JobId; })
+        .Optional(/*init*/ false);
+
+    registrar.ParameterWithUniversalAccessor<std::optional<TJobTraceId>>(
+        "trace_id",
+        [] (TThis* command) -> auto& {return command->Options.TraceId; })
+        .Optional(/*init*/ false);
+
+    registrar.ParameterWithUniversalAccessor<std::optional<i64>>(
+        "from_event_index",
+        [] (TThis* command) -> auto& {return command->Options.FromEventIndex; })
+        .Optional(/*init*/ false);
+
+    registrar.ParameterWithUniversalAccessor<std::optional<i64>>(
+        "to_event_index",
+        [] (TThis* command) -> auto& {return command->Options.ToEventIndex; })
+        .Optional(/*init*/ false);
+
+    registrar.ParameterWithUniversalAccessor<std::optional<i64>>(
+        "from_time",
+        [] (TThis* command) -> auto& {return command->Options.FromTime; })
+        .Optional(/*init*/ false);
+
+    registrar.ParameterWithUniversalAccessor<std::optional<i64>>(
+        "to_time",
+        [] (TThis* command) -> auto& {return command->Options.ToTime; })
+        .Optional(/*init*/ false);
+}
+
+void TGetJobTraceCommand::DoExecute(ICommandContextPtr context)
+{
+    auto result = WaitFor(context->GetClient()->GetJobTrace(OperationIdOrAlias, Options))
+        .ValueOrThrow();
+
+    context->ProduceOutputValue(BuildYsonStringFluently()
+        .DoListFor(result, [&] (TFluentList fluent, const TJobTraceEvent& traceEvent) {
+            Serialize(traceEvent, fluent.GetConsumer());
+        }));
+}
+
+////////////////////////////////////////////////////////////////////////////////
+
 void TGetJobFailContextCommand::Register(TRegistrar registrar)
 {
     registrar.Parameter("job_id", &TThis::JobId);
diff --git a/yt/yt/client/driver/scheduler_commands.h b/yt/yt/client/driver/scheduler_commands.h
index 280d55e331..59d1e9d21e 100644
--- a/yt/yt/client/driver/scheduler_commands.h
+++ b/yt/yt/client/driver/scheduler_commands.h
@@ -137,6 +137,20 @@ private:
 
 ////////////////////////////////////////////////////////////////////////////////
 
+class TGetJobTraceCommand
+    : public TSimpleOperationCommandBase<NApi::TGetJobTraceOptions>
+{
+public:
+    REGISTER_YSON_STRUCT_LITE(TGetJobTraceCommand);
+
+    static void Register(TRegistrar registrar);
+
+private:
+    void DoExecute(ICommandContextPtr context) override;
+};
+
+////////////////////////////////////////////////////////////////////////////////
+
 class TGetJobFailContextCommand
     : public TSimpleOperationCommandBase<NApi::TGetJobFailContextOptions>
 {
diff --git a/yt/yt/client/federated/client.cpp b/yt/yt/client/federated/client.cpp
index a3410ff50a..ff1b7505d6 100644
--- a/yt/yt/client/federated/client.cpp
+++ b/yt/yt/client/federated/client.cpp
@@ -407,6 +407,7 @@ public:
     UNIMPLEMENTED_METHOD(TFuture<NYson::TYsonString>, GetJobInputPaths, (NJobTrackerClient::TJobId, const TGetJobInputPathsOptions&));
     UNIMPLEMENTED_METHOD(TFuture<NYson::TYsonString>, GetJobSpec, (NJobTrackerClient::TJobId, const TGetJobSpecOptions&));
     UNIMPLEMENTED_METHOD(TFuture<TGetJobStderrResponse>, GetJobStderr, (const NScheduler::TOperationIdOrAlias&, NJobTrackerClient::TJobId, const TGetJobStderrOptions&));
+    UNIMPLEMENTED_METHOD(TFuture<std::vector<TJobTraceEvent>>, GetJobTrace, (const NScheduler::TOperationIdOrAlias&, const TGetJobTraceOptions&));
     UNIMPLEMENTED_METHOD(TFuture<TSharedRef>, GetJobFailContext, (const NScheduler::TOperationIdOrAlias&, NJobTrackerClient::TJobId, const TGetJobFailContextOptions&));
     UNIMPLEMENTED_METHOD(TFuture<TListOperationsResult>, ListOperations, (const TListOperationsOptions&));
     UNIMPLEMENTED_METHOD(TFuture<TListJobsResult>, ListJobs, (const NScheduler::TOperationIdOrAlias&, const TListJobsOptions&));
diff --git a/yt/yt/client/hedging/hedging.cpp b/yt/yt/client/hedging/hedging.cpp
index ae0376bcda..707a6759b8 100644
--- a/yt/yt/client/hedging/hedging.cpp
+++ b/yt/yt/client/hedging/hedging.cpp
@@ -164,6 +164,7 @@ public:
     UNSUPPORTED_METHOD(TFuture<NYson::TYsonString>, GetJobInputPaths, (NJobTrackerClient::TJobId, const TGetJobInputPathsOptions&));
     UNSUPPORTED_METHOD(TFuture<NYson::TYsonString>, GetJobSpec, (NJobTrackerClient::TJobId, const TGetJobSpecOptions&));
     UNSUPPORTED_METHOD(TFuture<TGetJobStderrResponse>, GetJobStderr, (const NScheduler::TOperationIdOrAlias&, NJobTrackerClient::TJobId, const TGetJobStderrOptions&));
+    UNSUPPORTED_METHOD(TFuture<std::vector<TJobTraceEvent>>, GetJobTrace, (const NScheduler::TOperationIdOrAlias&, const TGetJobTraceOptions&));
     UNSUPPORTED_METHOD(TFuture<TSharedRef>, GetJobFailContext, (const NScheduler::TOperationIdOrAlias&, NJobTrackerClient::TJobId, const TGetJobFailContextOptions&));
     UNSUPPORTED_METHOD(TFuture<TListOperationsResult>, ListOperations, (const TListOperationsOptions&));
     UNSUPPORTED_METHOD(TFuture<TListJobsResult>, ListJobs, (const NScheduler::TOperationIdOrAlias&, const TListJobsOptions&));
diff --git a/yt/yt/client/scheduler/public.h b/yt/yt/client/scheduler/public.h
index d92ae73f78..bcc8c61f1a 100644
--- a/yt/yt/client/scheduler/public.h
+++ b/yt/yt/client/scheduler/public.h
@@ -8,6 +8,12 @@ namespace NYT::NScheduler {
 
 ////////////////////////////////////////////////////////////////////////////////
 
+YT_DEFINE_STRONG_TYPEDEF(TJobTraceId, TGuid);
+
+extern const TJobTraceId NullJobTraceId;
+
+////////////////////////////////////////////////////////////////////////////////
+
 using NJobTrackerClient::TJobId;
 using NJobTrackerClient::TOperationId;
 
diff --git a/yt/yt/client/unittests/mock/client.h b/yt/yt/client/unittests/mock/client.h
index b8217c8657..e90e6b8e71 100644
--- a/yt/yt/client/unittests/mock/client.h
+++ b/yt/yt/client/unittests/mock/client.h
@@ -626,6 +626,11 @@ public:
         const TGetJobStderrOptions& options),
         (override));
 
+    MOCK_METHOD(TFuture<std::vector<TJobTraceEvent>>, GetJobTrace, (
+        const NScheduler::TOperationIdOrAlias& operationIdOrAlias,
+        const TGetJobTraceOptions& options),
+        (override));
+
     MOCK_METHOD(TFuture<TSharedRef>, GetJobFailContext, (
         const NScheduler::TOperationIdOrAlias& operationIdOrAlias,
         NJobTrackerClient::TJobId jobId,
diff --git a/yt/yt/core/ytree/tree_builder.h b/yt/yt/core/ytree/tree_builder.h
index 2af4d56a8d..fd959280d7 100644
--- a/yt/yt/core/ytree/tree_builder.h
+++ b/yt/yt/core/ytree/tree_builder.h
@@ -52,4 +52,3 @@ std::unique_ptr<ITreeBuilder> CreateBuilderFromFactory(INodeFactory* factory);
 ////////////////////////////////////////////////////////////////////////////////
 
 } // namespace NYT::NYTree
-
diff --git a/yt/yt_proto/yt/client/api/rpc_proxy/proto/api_service.proto b/yt/yt_proto/yt/client/api/rpc_proxy/proto/api_service.proto
index 7f3bae6c0a..0c7bc9eddc 100644
--- a/yt/yt_proto/yt/client/api/rpc_proxy/proto/api_service.proto
+++ b/yt/yt_proto/yt/client/api/rpc_proxy/proto/api_service.proto
@@ -2485,6 +2485,27 @@ message TRspGetJobStderr
 
 ////////////////////////////////////////////////////////////////////////////////
 
+message TReqGetJobTrace
+{
+    oneof operation_id_or_alias {
+        NYT.NProto.TGuid operation_id = 1;
+        string operation_alias = 2;
+    }
+    optional NYT.NProto.TGuid job_id = 3;
+    optional NYT.NProto.TGuid trace_id = 4;
+    optional int64 from_event_index = 5;
+    optional int64 to_event_index = 6;
+    optional int64 from_time = 7;
+    optional int64 to_time = 8;
+}
+
+message TRspGetJobTrace
+{
+    repeated TJobTraceEvent events = 1;
+}
+
+////////////////////////////////////////////////////////////////////////////////
+
 message TReqGetJobFailContext
 {
     oneof operation_id_or_alias {
@@ -3157,6 +3178,16 @@ message TListJobsResult
     repeated NYT.NProto.TError errors = 6;
 }
 
+message TJobTraceEvent
+{
+    required NYT.NProto.TGuid operation_id = 1;
+    required NYT.NProto.TGuid job_id = 2;
+    required NYT.NProto.TGuid trace_id = 3;
+    required int64 event_index = 4;
+    required string event = 5;
+    required int64 event_time = 6; // TInstant
+}
+
 ////////////////////////////////////////////////////////////////////////////////
 
 message TGetFileFromCacheResult
author	omgronny <omgronny@yandex-team.com>	2024-10-11 21:16:54 +0300
committer	omgronny <omgronny@yandex-team.com>	2024-10-11 21:28:33 +0300
commit	c7307fc40d61e2a996fd7c22ee9ef52663897b3b (patch)
tree	cf38637a1ef654e984c4847260253dd8195bff66 /yt
parent	1df197e6035ea9826bfedee7d48812e318ba9c7a (diff)
download	ydb-c7307fc40d61e2a996fd7c22ee9ef52663897b3b.tar.gz