Oracle Cloud Infrastructure Documentation

Metrics

Utilization Metrics

The following table shows the utilization metrics emitted by Ops Insights.

Metric Name Specific Dimensions Description
WarehouseCpuUtilization

resourceId - Ops Insights ID for the warehouse

resourceDisplayName - Display Name of the Ops Insights warehouse.

CPU Utilization of the ADW provisioned for the Ops Insights warehouse in percentage.
DaysToReachHighUtilization
resourceMetric -
  • Databases use: CPU, MEMORY, STORAGE
  • Hosts use: CPU, LOGICAL_MEMORY, STORAGE
  • Exadata use: CPU, STORAGE, MEMORY, IOPS, THROUGHPUT

aggregateDataMeasure - Indicates what underlying aggregate measure is being used in the forecast. Currently this can be AVG or MAX

forecastModel: Indicates which forecast model is being used in the forecast. Currently this can be SEASONALITY_AWARE, LINEAR_REGRESSION, or AUTOML

exceededForecastWindow - Indicates whether the number of days returned is equivalent to the amount of days being forecasted. This should be used in the alarms, like so: DaysToReachHighUtilization[1D]{resourceMetric="STORAGE", resourceType="Exadata", exceededForecastWindow="false"}.grouping(telemetrySource,resourceId).mean() < 30

Note

For additional dimensions for this metric, see Dimensions Common Across Ops Insights Metrics.
Days to reach high utilization (above default setting of 75%) for a given resource type and resource metric.

To modify utilization thresholds from the default settings see: Changing Utilization Thresholds.

DaysToReachLowUtilization
resourceMetric -
  • Databases use: CPU, MEMORY, STORAGE
  • Hosts use: CPU, LOGICAL_MEMORY, STORAGE
  • Exadata use: CPU, STORAGE, MEMORY, IOPS, THROUGHPUT

aggregateDataMeasure - Indicates what underlying aggregate measure is being used in the forecast. Currently this can be AVG or MAX

forecastModel: Indicates which forecast model is being used in the forecast. Currently this can be SEASONALITY_AWARE, LINEAR_REGRESSION, or AUTOML

exceededForecastWindow - Indicates whether the number of days returned is equivalent to the amount of days being forecasted. This should be used in the alarms, like so: DaysToReachHighUtilization[1D]{resourceMetric="STORAGE", resourceType="Exadata", exceededForecastWindow="false"}.grouping(telemetrySource,resourceId).mean() < 30

Note

For additional dimensions for this metric, see Dimensions Common Across Ops Insights Metrics.
Days to reach low utilization (below default setting of 25%) for a given resource type and resource metric.

To modify utilization thresholds from the default settings see: Changing Utilization Thresholds.

Data Flow Metric

Ops Insights consumes data coming from different types of sources such as Autonomous databases , Enterprise Manager targets (databases, hosts, Exadata, etc.) and Management Agent targets (external databases, hosts, etc). The data gap metric allows you to set up alarms in the event data from these sources has stopped for the last 1 or 2 days .

Note

For examples on setting up alarms for the data flow metric, see Specific Alarm Conditions (Data Flow Delays).
Metric Name Dimensions Description
DataFlowDelayInHrs

sourceIdentifier - This will be Enterprise Manager Bridge Id for Enterprise Manager target, agent Id for agent based target and OCID of ADW for Autonomous Database targets.

sourceEntityIdentifier - This will be the Enterprise Manager target GUID for Enterprise Manager target, Cloud Infrastructure database Id for Management Agent based targets.

associatedResourceId - This will only be populated for Autonomous Database targets and it will be the OCID of the Autonomous Database.

dataProcessingFrequencyInHrs - Frequency of data processing in hours

Note

For additional dimensions for this metric, see Dimensions Common Across Ops Insights Metrics.
Number of hours ago at which the data was last processed for given target and metric

Data Flow Metric Examples

The following table shows the possible values of the dataProcessingFrequencyInHrs dimension for different resource types.

dataProcessingFrequencyInHrs Value Resource Example telemetrySourceType Description
1.00 Enterprise Manager managed DB EnterpriseManager Loads every hour to process performance metric data accumulated in the Object Storage bucket for Enterprise Manager managed DB targets.
3.00

Autonomous DB

Database Cloud Service DB

Enterprise Manager managed DB

Enterprise Manager managed host

Exadata Cell

Cloud Infrastructure DB

CloudInfrastructure

EnterpriseManager

AgentService

Loads every 3 hrs to get hourly performance metric data from the Monitoring service (for Cloud Infrastructure and Autonomous DBs) , Object Storage bucket (for Enterprise Manager managed targets) or for generating hourly rollups from the raw data ingested via ingestion APIs.
12.00

Autonomous DB

Database Cloud Service DB

Enterprise Manager managed DB

Cloud Infrastructure DB

CloudInfrastructure

EnterpriseManager

AgentService

 Every 12 hours, load daily performance metrics data by reading raw data from the Ops Insights data store.
24.00

Enterprise Manager managed DB

Enterprise Manager managed host

Exadata Cell

 EnterpriseManager

Every 24 hrs, 2 ETLs are run to process data for Enterprise Manager managed targets.

One ETL to load daily performance metrics data from the Object Storage bucket for Enterprise Manager managed targets.

Another ETL to load configuration metrics data from the object storage bucket for Enterprise Manager managed targets.

Ops Insights Collection Issue Metric

The MetricCollectionErrors metric contains the number of collection errors for a given target and source metric name. This metric is currently only applicable for Oracle Cloud (ExaDB-D / Oracle Base Database / Autonomous) and Enterprise Manager based databases.

Metric Name Specific Dimensions Description
MetricCollectionErrors

associatedResourceId - This will be the DBAAS OCID for the resource.

sourceMetricName - Name of the metric collection which is failing.

This dimension can be one of the following:

  • ASHSqlStats
  • ASHSqlTexts
  • database_instance_load
  • memory_usage
  • storage_usage
  • tablespace_usage
  • db_external_properties
  • db_external_instance
  • db_os_config_instance
  • SqlPlan
  • SqlStats

ErrorCategory - DatabaseConnection/QueryExecution

Cause - <actual ORA error code if available> or NA

e.g., ORA-12850

Note

For additional dimensions for this metric, see Dimensions Common Across Ops Insights Metrics.
Number of collection errors for given target and source metric name.