Our docs are moving…

Check out our new Server Monitoring Doc.

Server Monitoring

Data garnered from logs can be used for more than just troubleshooting development issues.  Logs contain important statistics that can help system administrators determine whether a server is losing CPU cycles to a run away program or an errant process is chewing up available memory. Having a healthy system will allow for applications, platforms, and programs running smoothly.

More and more of our customers have started using their logs for server monitoring and performance management using the latest versions of our optional agents for Windows and Unix. Utilizing these latest builds you’ll now start to see server statistics come over into your Logentries account from your agent-based machines.After upgrading to the latest release of the agent, you’ll start to see a new log present in your hosts inside of Logentries.This log file will be named logentries-stats.log (see below).

Once the logs have populated within your environment, it’s easy to start to search and correlate information based on these statistics. The following system performance values are brought into the log:

CPU.system: The amount of CPU taken by system processes.
CPU.user: The amount of CPU being taken by user tasks.
CPU.idle: The amount of CPU that is not being used (idle).
Mem.total: Total memory available to the system in bytes.
Mem.active: Total memory being used in bytes.
Disk.write: Disk write I/O in bytes.
Disk.read: Disk read I/O in bytes.
Net.send: Network bytes sent.
Net.resv: Network bytes received.

Below is an example log displaying the information above:

As this information is brought in, customers can use our advanced search and analytics to diagnose system performance within their environment. Coupled with real-time alerts that can be scheduled based on those searches, customers will receive notification if their servers encounter a spike in CPU utilization or memory consumption.

For example, a search pattern of: 


would alert a customer when a server reached 80% CPU utilization.

Likewise, a search patter of:


would alert a customer when 90% of a 16gb memory computer was used up.

Adding in our Average, Sum, and Group-By functionality, you can plot out CPU, Network, Memory, or Disk I/O graphs.

Back to Top