All PingDirectory Servers have the capability to monitor the health of the server and host system
they run on for diagnostic review and troubleshooting. Initially, the servers do not collect any
performance data until they are prepared for monitoring by a Metrics Server using the
monitored-servers add-servers
tool or an administrator enables system health
data collection for real-time inspection and querying. At a high level, all of the important
server and machine metrics which can be monitored are available in the
cn=monitor
backend.
The Stats Collector plugin is the primary driver of performance data collection for LDAP, server response, replication, local JE databases, and host system machine metrics. Stats Collector configuration determines the sample and collection intervals, granularity of data (basic, extended, verbose), types of host system collection (cpu, disk, network) and what kind of data aggregation occurs for LDAP application statistics. The Stats Collector plugin ensures that a PingDataMetrics Server is able to gather all of the detailed data required for a comprehensive diagnostic review.
The Stats Collector plugin relies exclusively on entries in the cn=monitor
backend to sample data using LDAP queries. The Stats Collector plugin is the primary driver
of performance data collection for LDAP, server response, replication, local JE databases,
and host system machine metrics. Stats Collector configuration determines the sample
and collection intervals, granularity of data (basic, extended, verbose), types of host
system collection (cpu, disk, network) and the type of data aggregation that occurs
for LDAP application statistics. The Stats Collector plugin is configured with the
dsconfig
tool and collects data using LDAP queries. For example,
the --server-info:extended
option includes collection for the following:
- CPU
- JVM memory
- Memory
- Disk information
- Network information
Utilization metrics are gathered via externally invoked OS commands, such as
iostat
and netstat
, using platform-specific arguments and
version-specific output parsing.
Enabling the Host System monitor provider automatically gathers CPU and memory utilization but
only optionally gathers disk and network information. Disk and network interfaces are enumerated
in the configuration by device names (e.g., eth0
or lo), and by disk device
names (e.g., sd1, sdab, sda2, scsi0
).