This page contains the following status and performance information:
- Data Collection, shows if the data collection process is running.
- Interfaces, the number of switch/router interfaces being monitored.
- Samples, the rate at which the server is processing traffic samples.
- Bytes, the byte rate of measurement data arriving at the server. In order to prevent data loss, this value must be kept well below the capacity of the link attaching the server to the network.
- Lost, the number of traffic samples that were lost in the network. If this number is high, it might indicate that congestion in the switches, in a link between the server and the switches (see Bytes), or it might indicate that the server itself is unable to keep up with the arriving samples (see Bytes, CPU and Memory). High loss rates generally occur because of excessive sampling. Increasing the sampling rate will usually reduce the number of lost samples.
- Unreachable, the number of interfaces that the server wishes to collect data from, but which have become unreachable over the network. Click on the link to see which interfaces are no longer accessible.
- Other Manager, the number of interfaces that cannot be monitored because another collector is already monitoring them. Click on the link to see which interfaces are unavailable and who is using them.
- CPU, the load average on the CPU. This number indicates the average number of tasks ready to run at any instant. If the number is less than 1 then the CPU has spare capacity. The number may exceed 1 during periods when the server is busy handling queries, but a persistently high number might indicate that the server is becoming overloaded (check Lost to see if lack of CPU resources might be causing data loss).
- Memory, the memory utilization and amount of free memory. It is important to ensure that there is always spare memory capacity, insufficient memory will cause the server to slow down and cause data to be lost (see Lost).
- Disk, the disk utilization and amount of free disk space.
Ports
The UDP and TCP ports that is listening on for measurement data are listed here. If you intend to use one of these measurement protocols, then you must ensure that any firewall on the server is configured to allow connections from your network devices to these ports. If you are using devices configured automatically via SNMP using the sFlow MIB, then the ports listed as Other Ports must be open too.