Alertables
The tables below contain a guide of what fields are available for each of the different intervals and types.
Node Statistics¶
The following fields are applicable for both the per-minute and hourly node level data.
| Field | Type | Platform | Description |
|---|---|---|---|
| Timestamp | String | All | The date and time at for which the alert is generated. |
| Hostname | String | All | The hostname of a device is a human-readable name that is used to identify the device on the network, rather than using its IP address. |
| CPU (%) | Number | All | CPU utilization is a key performance metric used in system monitoring and analysis, as it provides insights into how effectively the CPU is being used to perform tasks. High CPU utilization may indicate that the CPU is working at or near its maximum capacity, which could potentially lead to performance degradation or system slowdowns. Low CPU utilization, on the other hand, may indicate that the CPU is underutilized and there may be room for optimization to improve system performance. |
| CPU Wait (%) | Number | All | CPU wait time is an important metric in system performance analysis, as it indicates the amount of time that a CPU is not being utilized efficiently. High CPU wait times can indicate performance bottlenecks, such as slow I/O operations, inefficient memory access, or contention for system resources. |
| Memory (MB) | Number | All | Memory (RAM) is used to temporarily store data and instructions that are being actively used or processed by the CPU of a system. RAM is a critical component of a computer system that plays a crucial role in enabling the CPU to quickly access and process data and instructions, facilitating the smooth operation of an operating system and running applications. |
| Buffer Cache Memory (MB) | Number | Linux only | The Linux buffer cache is a part of the Linux kernel's file system caching mechanism, which aims to improve I/O performance by caching frequently accessed data in memory. It is an important feature of the Linux file system that helps improve overall system performance by reducing the latency of data access and minimizing the amount of I/O operations that need to be performed on storage devices. |
| Swap (MB) | Number | All | Swap space is typically a portion of disk that is reserved for when a system runs out of physical memory. The disk space is used as an extension of the system's physical memory. The data that is swapped out to the swap space is stored in a special format called a swap file or swap partition, and it can be read back into RAM when it is needed again. |
| Total Memory (MB) | Number | All | Total Memory is the size, in Megabytes of the available Memory on the system. Please see Memory for a more detailed description of its purpose. |
Session Statistics¶
| Field | Type | Platform | Description |
|---|---|---|---|
| Timestamp | String | All | The date and time at for which the alert is generated. |
| Hostname | String | All | The hostname of a device is a human-readable name that is used to identify the device on the network, rather than using its IP address. |
| CPU (%) | Number | All | This metric is used to measure the amount of CPU resources that is being consumed by a process. It can help identify processes that are using excessive CPU resources or causing performance issues. CPU utilization at the session level is measured as a percentage of a single CPU core. This means that this number can go above 100%. |
| Memory (MB) | Number | All | This metric is used to measure the amount of RAM resources that are being consumed by a process, and can help identify processes that are using excessive memory resources or causing performance issues. |
| Data Read (MB) | Number | All | This metric is used to measure the amount of I/O resources that are being consumed by a process, and can help identify processes that are reading excessive amounts of data from storage devices or causing I/O bottlenecks. |
| Data Written (MB) | Number | Linux, Windows | This metric is used to measure the amount of I/O resources that are being consumed by a process, and can help identify processes that are reading excessive amounts of data from storage devices or causing I/O bottlenecks. |
| Job Name | String | All | The name of the session as it is registered within the application. |
| Log Lines | Number | All | Where a log file is associated with a process, the metric refers to the count of log messages that are recorded in a log file. Monitoring the number of log lines written in one minute can be useful for understanding the activity level of a system or application, as well as for troubleshooting issues or identifying patterns of behavior that may indicate a problem. |
| Major Faults | Number | Linux, Windows | This metric refers to the count of hard faults. A hard fault occurs when a process attempts to access a page of memory that is not currently resident in RAM and must be loaded from disk. When a major fault occurs, the operating system must retrieve the necessary data from disk and load it into memory before allowing the process to continue. Monitoring major faults can be useful for identifying potential performance bottlenecks or issues with memory usage. If a process is generating a large number of major faults, it may indicate that it is frequently accessing data that is not currently resident in RAM, which can slow down overall system performance. |
| Minor Faults | Number | Linux, Windows | This metric refers to the count of soft faults. A soft fault occurs when a process requests data from virtual memory that is not currently in physical memory, but is still present in the system's page cache or other buffers. Minor faults are a normal and expected part of the virtual memory management system used by modern operating systems, and are typically not a cause for concern. |
| Memory Share | Number | All | This metric refers to the amount of memory that is shared between a specific process and other processes running on a computer system. Shared memory allows processes to communicate and share data with each other without the need for inter-process communication (IPC) mechanisms such as message passing or file I/O. When a process uses shared memory, a portion of its memory space is mapped to a common memory region that can be accessed by other processes. This shared memory can be used to share data or communicate between processes in a fast and efficient manner. |
| Peak CPU (%) | Number | All | This metric refers to the maximum utilization of CPU for a process on a system in the given period of time. |
| Session Type | String | All | This is the type of session as categorized within the application. |
| Total Data Read (MB) | Number | All | This metric refers to the total amount of data read, to any device, by a process up to this period of time. |
| Total Data Written (MB) | Number | Linux, Windows | This metric refers to the total amount of data written, to any device, by a process during the specified period of time. |
| User CPU (%) | Number | Linux, Windows | This metric refers to the percentage of CPU that is being used by a process to execute user-level instructions. |
| System CPU (%) | Number | Linux, Windows | This metric refers to the percentage of CPU that is being for system-level purposes. System-level processes are typically those that are responsible for managing the system's resources and providing essential services to user-level processes. |
| UTILLOC Size (MB) | Number | All | This field is used to store two different numbers depending on the session's type. It can either be the size of SAS' UTILLOC which is a temporary location which enables threaded applications to store utility files. Or, it can be the size of the in-memory and disk-based CAS cache. |
| Virtual Memory (MB) | Number | All | Virtual memory is the amount of memory that is allocated to a specific process running on a computer system. Virtual memory is a technique used by operating systems to allow a process to use more memory than is physically available on the system by temporarily transferring data to and from disk storage. |
| WORK Size (MB) | Number | All | This field is used to store two different numbers depending on the session's type. It can either be the size of SAS' WORK location which is a temporary location used to store intermediate results during data processing. Or, it can be the size of the in-memory CAS cache. |
CAS Action Executions¶
| Field | Type | Platform | Description |
|---|---|---|---|
| Timestamp | String | All | The date and time at for which the alert is generated. |
| Hostname | String | All | The hostname of a device is a human-readable name that is used to identify the device on the network, rather than using its IP address. |
| Action Set | String | All | The action set is the parent group of a particular set of CAS actions. Examples include: clustering, deepLearn, regression, and svm |
| Action Name | String | All | The action name is the name of the CAS action that is being executed. Examples include: kClus, dlTrain, logistic, and svmTrain |
| Action Details | String | All | The details, if available of the action. This information may include options passed to the action. This data is returned as a string formatted as JSON. |
| Process PID | Number | All | The process ID of the CAS session where the action was executed. |
| Start Timestamp | String | All | The start time of the execution of the action. |
| End Timestamp | String | All | The end time of the execution of the action. |
| User | String | All | The user that executed the action. |
SAS Procedure Executions¶
| Field | Type | Platform | Description |
|---|---|---|---|
| Timestamp | String | All | The date and time at for which the alert is generated. |
| Hostname | String | All | The hostname of a device is a human-readable name that is used to identify the device on the network, rather than using its IP address. |
| Proc Name | String | All | The name of the SAS procedure that was executed. |
| Real Time | Number | All | The duration of the procedure execution |
| Process PID | Number | All | The process ID of the SAS session where the procedure was executed. |
| Start Timestamp | String | All | The start time of the execution of the procedure. |
| End Timestamp | String | All | The end time of the execution of the procedure. |
| User | String | All | The user that executed the procedure. |
Dataset Access Events¶
| Field | Type | Platform | Description |
|---|---|---|---|
| Timestamp | String | All | The date and time at for which the alert is generated. |
| Hostname | String | All | The hostname of a device is a human-readable name that is used to identify the device on the network, rather than using its IP address. |
| Access Time | String | All | The date and time at which the dataset was accessed by the session. |
| Engine | String | All | The SAS/ACCESS engine that was used to access the dataset. |
| Libref | String | All | In SAS, a libref is a shorthand name that is assigned to a SAS library. |
| Mode | String | All | The mode in which the dataset was accessed. Values are: READ, WRITE, and UPDATE. |
| Path | String | All | The location of the data. For file based data like SAS7BDAT file, this will be the folder in which the dataset is stored, for databases, this could be the IP address/hostname of the database server or the Data Source Name (DSN). |
| Process PID | String | All | The process ID of the SAS session which access the data |
| Table | String | All | The name of the table that was accessed. |
Log Events¶
| Field | Type | Platform | Description |
|---|---|---|---|
| Date & Time | String | All | The date and time at for which the alert is generated. |
| Hostname | String | All | The hostname of a device is a human-readable name that is used to identify the device on the network, rather than using its IP address. |
| Level | String | All | The level of detail or severity of the message generated. Values can be: trace, debug, info, warn, error, severe, and fatal. |
| Line Number | Number | All | The line number within the log file that the associated message is found. |
| Message | String | All | An entry that captures information about an event or activity that has occurred within an application or system. |
| Timestamp | String | All | The exact time that the message was written to the log. |
Session Summary Statistics¶
| Field | Type | Platform | Description |
|---|---|---|---|
| Date & Time | String | All | The date and time at for which the alert is generated. |
| Hostname | String | All | The hostname of a device is a human-readable name that is used to identify the device on the network, rather than using its IP address. |
| Average CPU (%) | Number | All | Measures the average CPU utilization for the lifetime of the session. |
| Average Memory (MB) | Number | All | Measure the average amount of memory (RAM) used during the lifetime of the session. |
| Container ID | String | All | For Kubernetes based deployments, this is the unique ID of the container in which the session was running. |
| Job Name | String | All | The name of the session as it appears in the application. |
| Namespace | String | All | For Kubernetes based deployments, this is the namespace in which the session (pod/container) ran. |
| Parent PID | Number | All | The parent process of the session. |
| Peak CPU (%) | Number | All | The maximum utilization of CPU for a session during its lifetime. |
| Peak Memory | Number | All | The maximum utilization of memory for a session during its lifetime. |
| Peak UTILLOC | Number | All | The maximum utilization of UTILLOC for a session during its lifetime. |
| Peak WORK | Number | All | The maximum utilization of WORK for a session during its lifetime. |
| Pod Name | String | All | For Kubernetes deployments, this is the unique name of the pod in which the session ran. |
| Priority | Number | All | The numerical representation of the priority with which the session was started. |
| Processor | Number | All | The processor ID of the processor on which the session was initially scheduled. |
| Queue | String | All | For SAS Grid deployments, this is the LSF queue that the session was executed in. |
| Session Type | String | All | The category of process that this session belongs to. |
| Total Data Read (MB) | Number | All | The total amount of data read, from any device, during the lifetime of the session. |
| Total Data Written (MB) | Number | All | The total amount of data written, to any device, during the lifetime of the session. |
| Total Duration (ms) | Number | All | The total duration of the session in milliseconds. |
| Total Log Lines | Number | All | The total number of log lines written during the lifetime of the session. |
| User | String | All | The username of the person that owned the session. |
Last update:
April 24, 2023
Created: April 24, 2023
Created: April 24, 2023