You can register Linux hosts and Amazon Elastic Compute Cloud (EC2) Linux instances to collect system health and performance metrics, such as CPU speed and available memory.
Note:
Each host must have its own installation of ArcGIS Monitor Agent. Once Monitor Agent is registered with Monitor Server, the host is automatically registered for monitoring.
Prerequisites for monitoring
The following prerequisites are necessary for monitoring Linux hosts:
- The Linux host that you want to register must be running a distribution and software version that is supported by ArcGIS Monitor.
- The sysstat package is required for monitoring.
Register Linux hosts for monitoring
You must be assigned the Administrator or Manager role to register components.
- Install Monitor Agent on the Linux host that you want to register.
- Start a terminal session on the Monitor Agent machine.
- Run the command below, editing the command line parameters to include the root URL to Monitor Server and your username and password.
./<Monitor Agent installation location>/arcgis-monitor-agent admin:system:register --server-url <https://monitorserver.domain.com:30443/arcgis> --username <user> --password <password>
The following is an example:
./usr/lib/arcgis/monitor/agent/arcgis-monitor-agent admin:system:register --server-url https://monitorserver.domain.com:30443/arcgis --username jdoe --password "my password"
The Monitor Agent instance is registered with Monitor Server. The host machine is automatically registered as a Linux host component and added to the list of monitored components on the Monitoring page.
Once a component is registered, you can configure alert rules and add the component to collections.
Note:
Other software component types, such as ArcGIS software and databases, are automatically discovered on the host machine and must be registered individually to enable monitoring.
Observers
The following table describes the observers that are responsible for collecting attribute and metric data:
Observer name | Description |
---|---|
Inventory | The Inventory observer scans for changes and manages attribute data across the Linux host. The default interval for this observer is 15 minutes; however, you can configure it to run at longer intervals since changes to the host's configuration are typically infrequent. |
Metrics | The Metrics observer collects Linux host metric data. The default interval for this observer is 1 minute to retrieve data frequently. It is recommended that you do not run this observer at longer intervals. |
Collected attributes
The subsections below describe the attributes that are collected for Linux hosts.
Host attributes
The following table lists the attributes that are collected for Linux hosts:
Attribute name | Description | Data type |
---|---|---|
Class | The instance type (virtual or physical) | String |
CPU cores logical | The CPU's total number of logical cores | Integer |
CPU cores physical | The CPU's total number of physical cores | Integer |
CPU name | The CPU name and model information | String |
CPU speed | The speed of the host's CPU | Float |
Memory total | The host's total amount of memory (RAM) | Float |
Network speed | The speed of the network | Integer |
Started at | The date and time the host was last started | Datetime |
UTC offset | The host's UTC time zone offset | Integer |
Version | The software version of the operating system | String |
Storage attributes
The following table lists the attributes collected for a Linux host's storage resources:
Attribute name | Description | Data type |
---|---|---|
Storage total | The size of the volume | Float |
Collected metrics
The subsections below describe the metrics that are collected for Linux hosts.
Host metrics
The following table lists the metrics collected for Linux hosts:
Metric name | Description | Data type | Unit |
---|---|---|---|
Connectivity | The agent's ability to successfully connect to the component. | Float | Percent |
CPU Cores Utilized | The number of logical CPU cores in use. | Float | Total |
CPU Utilized | The proportion of CPU resources in use. Values approaching 100 percent may indicate that more resources are needed to handle the workload. | Float | Percent |
Memory Available | The total amount of physical memory. | Float | Gigabytes |
Memory Used | The amount of memory in use. | Float | Gigabytes |
Memory Utilized | The proportion of memory in use. Values approaching 100 percent may indicate that more resources are needed to handle the workload. | Float | Percent |
Network Incoming | The rate of network data received. | Float | Megabits/second |
Network Incoming Utilized | The proportion of incoming network bandwidth in use. | Float | Percent |
Network Outgoing | The rate of network data sent. | Float | Megabits/second |
Network Outgoing Utilized | The proportion of outgoing network bandwidth in use. | Float | Percent |
Pagefile Available | The total pagefile or swap memory resources available. | Float | Gigabytes |
Pagefile Used | The total pagefile or swap memory resources in use. | Float | Gigabytes |
Pagefile Utilized | The proportion of pagefile or swap memory resources in use. | Float | Percent |
Process metrics
Monitor collects information about the following processes:
- arcgisserver
- arcsoc
- arcgisportal
- geoeventserver
- python
- sqlserver
- postgres
- oracle
The following table lists the metrics collected for a Linux host's processes:
Metric name | Description | Data type | Unit |
---|---|---|---|
Process CPU Utilized | The proportion of CPU resources in use by the process. Values approaching 100 percent may indicate that more resources are needed to handle the workload. | Float | Percent |
Process Instances | The total number of process instances that are running. | Float | Total |
Process Memory Used | The amount of memory resources in use by the process. | Float | Gigabytes |
Process Memory Utilized | The proportion of memory in use by the process. Values approaching 100 percent may indicate that more resources are needed to handle the workload. | Float | Percent |
Storage metrics
The following table lists the metrics collected for a Linux host's storage resources:
Metric name | Description | Data type | Unit |
---|---|---|---|
Storage Capacity Available | The amount of available storage space | Float | Gigabytes |
Storage Capacity Used | The amount of storage space in use | Float | Gigabytes |
Storage Capacity Utilized | The proportion of storage capacity in use | Float | Percent |
Storage Read Rate | The speed at which data is read from the disk | Float | Gigabytes/second |
Storage Write Rate | The speed at which data is written to the disk | Float | Gigabytes/second |
Default alert rules
The following table lists the default alert rules that are configured for Linux hosts:
Metric name | Aggregation | Operator | Info threshold | Warning threshold | Critical threshold | Samples |
---|---|---|---|---|---|---|
Connectivity | Average | Is less than | — | — | 100 | 2 |
CPU Utilized | Average | Is greater than | — | 75 | 90 | 3 |
Memory Utilized | Average | Is greater than | — | 75 | 90 | 3 |
Pagefile Utilized | Average | Is greater than | — | 75 | 90 | 3 |
Process CPU Utilized | Average | Is greater than | — | 75 | 90 | 3 |
Process Memory Utilized | Average | Is greater than | — | 75 | 90 | 3 |
Storage Capacity Utilized | Average | Is greater than | — | 75 | 90 | 3 |
Note:
The default alert rules in the table above are the rules that are configured for components when Monitor is first deployed. If your Monitor administrator has configured default alert rules for the Monitor deployment, the values for the rules above may vary.