Skip to content

DPL: send error count in a delayed manner #5371

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

Merged
ktf merged 1 commit into from
Feb 3, 2021
Merged

DPL: send error count in a delayed manner #5371

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
2 changes: 0 additions & 2 deletions Framework/Core/include/Framework/DataProcessingDevice.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -74,7 +74,6 @@ struct DataProcessorContext {
AlgorithmSpec::ErrorCallback* error = nullptr;

std::function<void(o2::framework::RuntimeErrorRef e, InputRecord& record)>* errorHandling = nullptr;
int* errorCount = nullptr;
};

/// A device actually carrying out all the DPL
Expand Down Expand Up @@ -124,7 +123,6 @@ class DataProcessingDevice : public FairMQDevice
/// Completed actions
std::vector<DataRelayer::RecordAction> mCompleted;

int mErrorCount;
uint64_t mLastSlowMetricSentTimestamp = 0; /// The timestamp of the last time we sent slow metrics
uint64_t mLastMetricFlushedTimestamp = 0; /// The timestamp of the last time we actually flushed metrics
uint64_t mBeginIterationTimestamp = 0; /// The timestamp of when the current ConditionalRun was started
Expand Down
1 change: 1 addition & 0 deletions Framework/Core/include/Framework/DataProcessingStats.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,7 @@ struct DataProcessingStats {
int minLatency = 0;
int maxLatency = 0;
};
std::atomic<int> errorCount = 0;
std::atomic<int> pendingInputs = 0;
std::atomic<int> incomplete = 0;
std::atomic<int> inputParts = 0;
Expand Down
1 change: 1 addition & 0 deletions Framework/Core/src/CommonServices.cxx
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -480,6 +480,7 @@ auto sendRelayerMetrics(ServiceRegistry& registry, DataProcessingStats& stats) -
monitoring.send(Metric{(int)relayerStats.droppedIncomingMessages, "dropped_incoming_messages"}.addTag(Key::Subsystem, Value::DPL));
monitoring.send(Metric{(int)relayerStats.relayedMessages, "relayed_messages"}.addTag(Key::Subsystem, Value::DPL));

monitoring.send(Metric{(int)stats.errorCount, "errors"}.addTag(Key::Subsystem, Value::DPL));
monitoring.send(Metric{(int)stats.pendingInputs, "inputs/relayed/pending"}.addTag(Key::Subsystem, Value::DPL));
monitoring.send(Metric{(int)stats.incomplete, "inputs/relayed/incomplete"}.addTag(Key::Subsystem, Value::DPL));
monitoring.send(Metric{(int)stats.inputParts, "inputs/relayed/total"}.addTag(Key::Subsystem, Value::DPL));
Expand Down
17 changes: 6 additions & 11 deletions Framework/Core/src/DataProcessingDevice.cxx
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -94,8 +94,7 @@ DataProcessingDevice::DataProcessingDevice(DeviceSpec const& spec, ServiceRegist
mError{spec.algorithm.onError},
mConfigRegistry{nullptr},
mAllocator{&mTimingInfo, &registry, spec.outputs},
mServiceRegistry{registry},
mErrorCount{0}
mServiceRegistry{registry}
{
/// FIXME: move erro handling to a service?
if (mError != nullptr) {
Expand Down Expand Up @@ -427,7 +426,6 @@ void DataProcessingDevice::fillContext(DataProcessorContext& context)
context.error = &mError;
/// Callback for the error handling
context.errorHandling = &mErrorHandling;
context.errorCount = &mErrorCount;
}

void DataProcessingDevice::PreRun()
Expand Down Expand Up @@ -684,9 +682,8 @@ void DataProcessingDevice::handleData(DataProcessorContext& context, FairMQParts
return results;
};

auto reportError = [& registry = *context.registry, &context](const char* message) {
context.errorCount++;
registry.get<Monitoring>().send(Metric{*context.errorCount, "errors"}.addTag(Key::Subsystem, Value::DPL));
auto reportError = [&registry = *context.registry, &context](const char* message) {
registry.get<DataProcessingStats>().errorCount++;
};

auto handleValidMessages = [&parts, &context = context, &relayer = *context.relayer, &reportError](std::vector<InputType> const& types) {
Expand Down Expand Up @@ -782,9 +779,8 @@ bool DataProcessingDevice::tryDispatchComputation(DataProcessorContext& context,
// should work just fine.
std::vector<MessageSet> currentSetOfInputs;

auto reportError = [& registry = *context.registry, &context](const char* message) {
context.errorCount++;
registry.get<Monitoring>().send(Metric{*context.errorCount, "errors"}.addTag(Key::Subsystem, Value::DPL));
auto reportError = [&registry = *context.registry, &context](const char* message) {
registry.get<DataProcessingStats>().errorCount++;
};

// For the moment we have a simple "immediately dispatch" policy for stuff
Expand Down Expand Up @@ -1077,8 +1073,7 @@ bool DataProcessingDevice::tryDispatchComputation(DataProcessorContext& context,
void DataProcessingDevice::error(const char* msg)
{
LOG(ERROR) << msg;
mErrorCount++;
mServiceRegistry.get<Monitoring>().send(Metric{mErrorCount, "errors"}.addTag(Key::Subsystem, Value::DPL));
mServiceRegistry.get<DataProcessingStats>().errorCount++;
}

} // namespace o2::framework