The MonitoringManager is responsible for managing all global metrics of all nodes in the topology. More...

#include <MonitoringManager.hpp>

Public Member Functions
	MonitoringManager (TopologyPtr topology, RequestHandlerServicePtr requestHandlerService, Catalogs::Query::QueryCatalogPtr queryCatalog, MetricStorePtr metricStore, bool enableMonitoring)

	MonitoringManager (TopologyPtr topology, RequestHandlerServicePtr requestHandlerService, Catalogs::Query::QueryCatalogPtr queryCatalog, bool enableMonitoring)
	Ctor to create a MonitoringManger for a given topology. For communication the manager will use the corresponding RPC client. More...

	MonitoringManager (TopologyPtr topology, RequestHandlerServicePtr requestHandlerService, Catalogs::Query::QueryCatalogPtr queryCatalog)

	MonitoringManager (const MonitoringManager &)=default

	MonitoringManager (MonitoringManager &&)=default

MonitoringManager &	operator= (const MonitoringManager &)=default

MonitoringManager &	operator= (MonitoringManager &&)=default

	~MonitoringManager ()

bool	registerRemoteMonitoringPlans (const std::vector< WorkerId > &nodeIds, MonitoringPlanPtr monitoringPlan)
	Register a monitoring plan for given nodes. More...

nlohmann::json	requestRemoteMonitoringData (WorkerId nodeId)
	Get the monitoring data for a given node. Note: Multiple nodes are not possible, as every node can have a different monitoring plan and TupleBuffer is not supporting different nested schemas. More...

StoredNodeMetricsPtr	getMonitoringDataFromMetricStore (WorkerId nodeId)
	Requests monitoring data from metric store. More...

void	addMonitoringData (WorkerId nodeId, MetricPtr metrics)
	Receive arbitrary monitoring data from a given node. More...

void	removeMonitoringNode (WorkerId nodeId)
	Remove node from monitoring store. More...

MonitoringPlanPtr	getMonitoringPlan (WorkerId nodeId)
	Get the monitoring plan for a given node ID. If the node exists in the topology but has not a registered plan, MonitoringPlan::Default will be returned. If the node does not exist an NES exception is thrown. More...

bool	registerLogicalMonitoringStreams (const NES::Configurations::CoordinatorConfigurationPtr config)
	Registers the logical monitoring streams at the coordinator. More...

std::unordered_map< std::string, QueryId >	startOrRedeployMonitoringQueries (bool sync)
	Starts or redeploys monitoring queries at the coordinator. More...

QueryId	startOrRedeployMonitoringQuery (std::string monitoringStream, bool sync)
	Starts or redeploys monitoring queries at the coordinator. More...

bool	isMonitoringStream (std::string streamName) const
	Checks if the logical stream is a monitoring stream. More...

bool	stopRunningMonitoringQuery (std::string streamName, bool sync)
	Stops a given running monitoring query;. More...

bool	stopRunningMonitoringQueries (bool sync)
	Stops all running monitoring queries;. More...

MetricStorePtr	getMetricStore ()

const std::unordered_map< std::string, QueryId > &	getDeployedMonitoringQueries () const
	Get the deployed monitoring queries. More...

Detailed Description

The MonitoringManager is responsible for managing all global metrics of all nodes in the topology.

Constructor & Destructor Documentation

◆ MonitoringManager() [1/5]

NES::Monitoring::MonitoringManager::MonitoringManager	(	TopologyPtr	topology,
		RequestHandlerServicePtr	requestHandlerService,
		Catalogs::Query::QueryCatalogPtr	queryCatalog,
		MetricStorePtr	metricStore,
		bool	enableMonitoring
	)

Ctor to create a MonitoringManger for a given topology. For communication the manager will use the corresponding RPC client.

Parameters

topology	the topology
requestHandlerService	request handler
queryCatalog	the query catalog
metricStore	the metric store
enableMonitoring	flag to indicate if monitoring is enabled or not

References magic_enum::enum_name(), and NES_DEBUG.

Here is the call graph for this function:

◆ MonitoringManager() [2/5]

NES::Monitoring::MonitoringManager::MonitoringManager	(	TopologyPtr	topology,
		RequestHandlerServicePtr	requestHandlerService,
		Catalogs::Query::QueryCatalogPtr	queryCatalog,
		bool	enableMonitoring
	)

Ctor to create a MonitoringManger for a given topology. For communication the manager will use the corresponding RPC client.

Parameters

topology	the topology
requestHandlerService	request handler
queryCatalog	the query catalog
enableMonitoring	flag to indicate if monitoring is enabled or not

◆ MonitoringManager() [3/5]

NES::Monitoring::MonitoringManager::MonitoringManager	(	TopologyPtr	topology,
		RequestHandlerServicePtr	requestHandlerService,
		Catalogs::Query::QueryCatalogPtr	queryCatalog
	)

Ctor to create a MonitoringManger for a given topology. For communication the manager will use the corresponding RPC client.

Parameters

topology	the topology
requestHandlerService	the query service
queryCatalog	the query catalog

◆ MonitoringManager() [4/5]

NES::Monitoring::MonitoringManager::MonitoringManager ( const MonitoringManager & )

default

◆ MonitoringManager() [5/5]

NES::Monitoring::MonitoringManager::MonitoringManager ( MonitoringManager && )

default

◆ ~MonitoringManager()

NES::Monitoring::MonitoringManager::~MonitoringManager ( )

References NES_DEBUG.

Member Function Documentation

◆ addMonitoringData()

void NES::Monitoring::MonitoringManager::addMonitoringData	(	WorkerId	nodeId,
		MetricPtr	metrics
	)

Receive arbitrary monitoring data from a given node.

Parameters

nodeId
GroupedMetricValuesPtr	the grouped metric values

References magic_enum::enum_name(), NES_TRACE, and nodeId.

Here is the call graph for this function:

◆ getDeployedMonitoringQueries()

const std::unordered_map< std::string, QueryId > & NES::Monitoring::MonitoringManager::getDeployedMonitoringQueries ( ) const

Get the deployed monitoring queries.

Returns: A map logicalStreamName -> QueryId

◆ getMetricStore()

MetricStorePtr NES::Monitoring::MonitoringManager::getMetricStore ( )

Getter for the metric store

Returns: the metric store

◆ getMonitoringDataFromMetricStore()

StoredNodeMetricsPtr NES::Monitoring::MonitoringManager::getMonitoringDataFromMetricStore ( WorkerId nodeId )

Requests monitoring data from metric store.

Parameters

nodeId

Returns: the grouped metric values

◆ getMonitoringPlan()

MonitoringPlanPtr NES::Monitoring::MonitoringManager::getMonitoringPlan ( WorkerId nodeId )

Get the monitoring plan for a given node ID. If the node exists in the topology but has not a registered plan, MonitoringPlan::Default will be returned. If the node does not exist an NES exception is thrown.

Parameters

nodeId

Returns: The monitoring plan

References NES::Monitoring::MonitoringPlan::defaultPlan(), NES_DEBUG, NES_THROW_RUNTIME_ERROR, and nodeId.

Referenced by requestRemoteMonitoringData().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ isMonitoringStream()

bool NES::Monitoring::MonitoringManager::isMonitoringStream ( std::string streamName ) const

Checks if the logical stream is a monitoring stream.

Parameters

streamName

Returns: true if monitoring stream, else false

◆ operator=() [1/2]

MonitoringManager& NES::Monitoring::MonitoringManager::operator= ( const MonitoringManager & )

default

◆ operator=() [2/2]

MonitoringManager& NES::Monitoring::MonitoringManager::operator= ( MonitoringManager && )

default

◆ registerLogicalMonitoringStreams()

bool NES::Monitoring::MonitoringManager::registerLogicalMonitoringStreams ( const NES::Configurations::CoordinatorConfigurationPtr config )

Registers the logical monitoring streams at the coordinator.

Returns: true if monitoring is disabled or if the streams have been registered successfully, else false

References NES::Monitoring::MetricUtils::createMetricFromCollectorType(), magic_enum::enum_name(), NES::Monitoring::MetricUtils::getSchemaFromCollectorType(), NES::TestUtils::logicalSourceName(), NES_INFO, and NES_WARNING.

Here is the call graph for this function:

◆ registerRemoteMonitoringPlans()

bool NES::Monitoring::MonitoringManager::registerRemoteMonitoringPlans	(	const std::vector< WorkerId > &	nodeIds,
		MonitoringPlanPtr	monitoringPlan
	)

Register a monitoring plan for given nodes.

Parameters

nodeId
monitoringPlan

Returns: True, if successful, else false

References NES_DEBUG, NES_ERROR, and nodeId.

◆ removeMonitoringNode()

void NES::Monitoring::MonitoringManager::removeMonitoringNode ( WorkerId nodeId )

Remove node from monitoring store.

Parameters

nodeId

References NES_DEBUG, and nodeId.

◆ requestRemoteMonitoringData()

nlohmann::json NES::Monitoring::MonitoringManager::requestRemoteMonitoringData ( WorkerId nodeId )

Get the monitoring data for a given node. Note: Multiple nodes are not possible, as every node can have a different monitoring plan and TupleBuffer is not supporting different nested schemas.

Parameters

nodeId
tupleBuffer

Returns: the grouped metric values

References getMonitoringPlan(), NES_DEBUG, NES_ERROR, NES_THROW_RUNTIME_ERROR, and nodeId.

Here is the call graph for this function:

◆ startOrRedeployMonitoringQueries()

std::unordered_map< std::string, QueryId > NES::Monitoring::MonitoringManager::startOrRedeployMonitoringQueries ( bool sync )

Starts or redeploys monitoring queries at the coordinator.

Returns: true if successful, else false

References NES_ERROR, startOrRedeployMonitoringQuery(), and stopRunningMonitoringQuery().

Here is the call graph for this function:

◆ startOrRedeployMonitoringQuery()

QueryId NES::Monitoring::MonitoringManager::startOrRedeployMonitoringQuery	(	std::string	monitoringStream,
		bool	sync
	)

Starts or redeploys monitoring queries at the coordinator.

Parameters

the	logical stream name of the monitoring stream
bool	true if it should block, else false

Returns: the QueryID the of the monitoring stream

References NES::Optimizer::BottomUp, NES::Monitoring::MetricUtils::createCollectorTypeFromMetricType(), magic_enum::enum_name(), NES_ERROR, NES_INFO, and stopRunningMonitoringQuery().

Referenced by startOrRedeployMonitoringQueries().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ stopRunningMonitoringQueries()

bool NES::Monitoring::MonitoringManager::stopRunningMonitoringQueries ( bool sync )

Stops all running monitoring queries;.

Parameters

sync	if to true then block

Returns: true if success

References NES_ERROR, and stopRunningMonitoringQuery().

Here is the call graph for this function:

◆ stopRunningMonitoringQuery()

bool NES::Monitoring::MonitoringManager::stopRunningMonitoringQuery	(	std::string	streamName,
		bool	sync
	)

Stops a given running monitoring query;.

Parameters

the	name of the monitoring stream
sync	if it should block or not

Returns: true if success

References NES_ERROR, and NES_INFO.

Referenced by startOrRedeployMonitoringQueries(), startOrRedeployMonitoringQuery(), and stopRunningMonitoringQueries().

Here is the caller graph for this function:

The documentation for this class was generated from the following files:

/nebulastream/nes-coordinator/include/Monitoring/MonitoringManager.hpp
/nebulastream/nes-coordinator/src/Monitoring/MonitoringManager.cpp

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ MonitoringManager() [1/5]

◆ MonitoringManager() [2/5]

◆ MonitoringManager() [3/5]

◆ MonitoringManager() [4/5]

◆ MonitoringManager() [5/5]

◆ ~MonitoringManager()

Member Function Documentation

◆ addMonitoringData()

◆ getDeployedMonitoringQueries()

◆ getMetricStore()

◆ getMonitoringDataFromMetricStore()

◆ getMonitoringPlan()

◆ isMonitoringStream()

◆ operator=() [1/2]

◆ operator=() [2/2]

◆ registerLogicalMonitoringStreams()

◆ registerRemoteMonitoringPlans()

◆ removeMonitoringNode()

◆ requestRemoteMonitoringData()

◆ startOrRedeployMonitoringQueries()

◆ startOrRedeployMonitoringQuery()

◆ stopRunningMonitoringQueries()

◆ stopRunningMonitoringQuery()