Skip to content
share library_books

Sematext Process Monitoring

Process Monitoring gives you visibility into processes that use the most CPU or memory, regardless of whether they are running inside containers, in the cloud, your own VMs or bare-metal servers.

Process Monitoring is available in the Infrastructure Monitoring section of your Sematext Cloud account - your central place for hosts, virtual machines, containers, and process information.

Processes View

The processes view provides similar information as the Linux top command. It shows:

  • Top N processes in your entire infrastructure
  • Grouping metrics by hosts, process names, containers, etc.
  • Filtering by tags

The real-time view of the top N processes contains a time-series chart, a tile map, and a list with the process details that help you identify processes with the highest resource usage.

Expanding a row in the process list shows detailed information about the process:

  • PID
  • User
  • Group
  • Process name
  • Command
  • Start time
  • Umask
  • Gid
  • Tid
  • Tags like container id, container name etc.

Also, a real-time view of all process metrics is available in the metrics tab. This includes:

  • CPU usage
  • RSS memory
  • IO read/write bytes
  • IO read/write operations

How does it Work?

The Sematext Agent tracks all processes and ships the metrics of the top N processes. The agent tags all metrics with process and container metadata and ships the metrics to your Infra App.

Sematext UI provides visualizations, grouping and filter functionality.

Enabling Process Monitoring

Process Monitoring is enabled by default in the Sematext Agent configuration. In case you want to disable Process Monitoring set the following option in the config file:

process:
 enabled: false

After that the Sematext Agent needs to be restarted by running the following command:

service spm-monitor-st-agent restart

Gathered Data

The Sematext Agent gathers the following data about the process:

  • CPU usage
  • RSS memory
  • IO read/write bytes
  • IO read/write operations
  • Process metadata
  • Container metadata

Solving Problems With Process Monitoring

Here are some of the common use cases and issues that Process Monitoring helps to solve:

  • Real-time process monitoring
  • Extended visibility for at least 30 minutes
  • Save time to identify processes with the highest resource in your entire infrastructure
  • Replace time-consuming operations to check system resources with ssh and top Linux commands
  • Identify processes, containers or service with high resource usage