About Process Monitoring
Process monitoring allows you setting CPU, RAM and Virtual Memory thresholds for a process running on your Windows or Linux machine so that you get alerted if utilization of any of these resources by the process reaches some critical level preset by you.
Adding Process Monitoring
Adding Process Monitoring for Windows
Go to the Monitors menu, select Server/Device Monitors->Process and click Add.
The Add Server/Device Monitor – Step 1 window will open.
If you have already downloaded and installed Smart Agent for Windows, select it from the Agent list.
Make sure that your Smart Agent for Windows is running, as shown under Agent Status. If the Agent is stopped, run it from the Smart Agent Controller window (see Downloading, Installing and Running Smart Agent for Windows).
If you haven’t downloaded Smart Agent for Windows yet, click on the Download Agent button to download it. Follow the instructions under Downloading, Installing and Running Smart Agent for Windows.
Note that it may take up to 5 min for newly installed Agent to appear in the agent list (you can click Refresh to refresh the agent list).
Select the Windows Agent from the list, and click Next.
The Add Server/Device Monitor – Step 2 window will appear.
Select monitor group for this monitor from the Monitor Group combo box .
You can also create a new monitor group by clicking on the “+” icon next to Add Group.
Select the processes to monitor. For each selected process a separate Process monitor will be added in your dashboard.
You can configure here a threshold of Critical severity level for the new process monitor(s) (see Thresholds). You can also edit thresholds of the monitor after it is added in your dashboard.
The threshold you define here applies to all cores.
The number of consecutive checks the threshold condition must be met in order for the monitor to enter Critical state is set to be 1 in the threshold.
Enter the values under CPU Usage, Memory Usage and Virtual Memory Usage. These limits will apply to all the processes that you have selected here. Your new process monitor(s) will enter Critical state whenever any of these limits are crossed. You can then assign alert rules to your monitor(s) to start receiving alerts whenever the monitor(s) enters a problem state (see Alert Rules in Alerts 2.0).
To add more Server-Device monitors of different types to the same Smart Agent, click the green “+” button under Add More Monitors (see Adding Multiple Server-Device and Application Monitors for the Same Agent).
Click Add when finished.
The Alert Configuration window will appear.
- Click Send to All Contacts if you want to send alerts for this monitor to all contacts.
- Click Custom Configuration (Advanced) to configure sending alerts for this monitor only to selected contacts. The Add Alert Rules window will open. Click on the Add Alert Rule button to configure alert rule(s) for your monitor.
- Set up the alert rule and click Add Rule. See Alert Rules in Alerts 2.0 for more info about alert rules.
Note: To be able to add alert rules, you need to set up thresholds (Critical or Warning) in your monitor first.
A threshold includes two conditions connected by a logical OR. If any of the conditions matches the monitor will enter your specified state.
- Check fails (response timed out, connection failure, process not running, etc)
- Any or all of your selected metrics meet the failure conditions you specified for them
Note that you have a possibility to set threshold for the running processes only by not including the availability part (“Result is error”) in the threshold condition.
See Thresholds for more info.
- Click Don’t Send Alerts if you don’t want any alerts for this monitor to be sent.
Your newly created Server-Device monitor will appear in your Monitis dashboard. Wait until it loads the first results.
Adding Process Monitoring for Linux
Go to the Monitors menu, select Server/Device Monitors->Process and click Add.
The Add Server/Device Monitor – Step 1 window will open.
If you have already downloaded and installed Smart Agent for Linux, select it from the Agent list.
Make sure that your Smart Agent for Linux is running, as shown under Agent Status. If the Agent is stopped, run it on your Linux machine (see Downloading, Installing and Running Smart Agent for Linux).
If you haven’t downloaded Smart Agent for Linux yet, click on the Download Agent button to download it. Follow the instructions under Downloading, Installing and Running Smart Agent for Linux.
Note that it may take up to 5 min for newly installed Agent to appear in the list (you can click Refresh to refresh the agent list).
Select the Linux Agent from the list, and click Next.
The Add Server/Device Monitor – Step 2 window will appear.
Select monitor group for this monitor from the Monitor Group combo box.
You can also create a new monitor group by clicking on the “+” icon next to Add Group.
Select the processes to monitor. For each selected process a separate Process monitor will be added in your dashboard.
You can configure here a threshold of Critical severity level for the new process monitor(s) (see Thresholds). You can also edit thresholds of the monitor after it is added in your dashboard.
The threshold you define here applies to all cores.
The number of consecutive checks the threshold condition must be met in order for the monitor to enter Critical state is set to be 1 in the threshold.
Enter the values under CPU Usage, Memory Usage and Virtual Memory Usage. These limits will apply to all the processes that you have selected here. Your new process monitor(s) will enter Critical state whenever any of these limits are crossed. You can then assign alert rules to your monitor(s) to start receiving alerts whenever the monitor(s) enters a problem state (see Alert Rules in Alerts 2.0).
To add more Server-Device monitors of different types to the same Smart Agent, click the green “+” button under Add More Monitors (see Adding Multiple Server-Device and Application Monitors for the Same Agent).
Click Add when finished.
The Alert Configuration window will appear.
- Click Send to All Contacts if you want to send alerts for this monitor to all contacts.
- Click Custom Configuration (Advanced) to configure sending alerts for this monitor only to selected contacts. The Add Alert Rules window will open. Click on the Add Alert Rule button to configure alert rule(s) for your monitor.
- Set up the alert rule and click Add Rule. See Alert Rules in Alerts 2.0 for more info about alert rules.
Note: To be able to add alert rules, you need to set up thresholds (Critical or Warning) in your monitor first.
A threshold includes two conditions connected by a logical OR. If any of the conditions matches the monitor will enter your specified state.
- Check fails (response timed out, connection failure, process not running, etc)
- Any or all of your selected metrics meet the failure conditions you specified for them
Note that you have a possibility to set threshold for the running processes only by not including the availability part (“Result is error”) in the threshold condition.
See Thresholds for more info.
- Click Don’t Send Alerts if you don’t want any alerts for this monitor to be sent.
Your newly created Server-Device monitor will appear in your Monitis dashboard. Wait until it loads the first results.
Data Aggregation
Process monitor is a multi-value monitor, meaning that every monitoring check returns values for more than one object – in this case for every Process ID (PID) of the monitored process.
Depending on the selected time range, data in your Process monitor will be aggregated across PIDs as per the below rules.
If the Selected Time Range is Less Than or Equal to 24 Hours
If you select one of the following time ranges, which are all within the 24 hours interval:
- Today
- Last 24 hours
- Specific date
Data in your Process monitor will be aggregated as follows:
- In table view:
Data is not aggregated. You will be able to see metric values, e.g. CPU, Memory and Virtual Memory returned by every monitoring check for all the PIDs of the monitored process, as in the below screen of Process monitor for Linux with the selected time range Today.
- In chart view:
Data is aggregated across PIDs. For each check, the values of the metric returned for all PIDs are summed up and shown on the chart as the aggregated result for that check.
E.g.:
CPU_CHECK1 = SUM(CPU_PID1_CHECK1, CPU_PID2_CHECK1, …………………CPU_PIDn_CHECK1)
.
.
.
CPU_CHECKN = SUM(CPU_PID1_CHECKN, CPU_PID2_CHECKN, …………………CPU_PIDn_CHECKN)
The below screen shows chart view of the same Process monitor for Linux for the same time range Today.
The shown value for CPU is the sum of CPU metric’s values returned for all PIDs for that check.
If the Selected Time Range is More Than 24 Hours
If you select one of the following time ranges, which are all larger than the 24 hours interval:
- 3 days
- 7 days
- 30 days
- Date Range
Data in your Process monitor will be aggregated as follows:
- Both in chart and table view:
Data is aggregated across PIDs.
For every reporting interval (1 hour for the 3 days time range, 1 day for 7 days, 30 days and Custom date range):
- For each check, the values of the metric returned for all PIDs are summed up.
E.g.:
CPU_CHECK1 = SUM(CPU_PID1_CHECK1, CPU_PID2_CHECK1, …………………CPU_PIDn_CHECK1)
.
.
.
CPU_CHECKN = SUM(CPU_PID1_CHECKN, CPU_PID2_CHECKN, …………………CPU_PIDn_CHECKN)
- Data is then aggregated using the aggregation function selected in the monitor’s settings.
See Data Aggregation in Monitors for more info.
For example, if Average is selected as the aggregation function:
The average of the sums across all checks is calculated and reported as the result for the interval.
E.g.:
CPU_INTERVAL1 = AVERAGE (SUM(CPU_PID1_CHECK1, CPU_PID2_CHECK1, ………CPU_PIDn_CHECK1), …………………….. SUM(CPU_PID1_CHECKN, CPU_PID2_CHECKN, ………CPU_PIDn_CHECKN))
As you can see it both in the table and chart views of the same monitor, when e.g. 3 days is selected as time range, the monitor shows one (aggregated) value per each reporting interval (1 hour in case of the selected 3 days time range).