PathSolutions is designed to provide visibility into all links in a network. The benefit of this is that errors and bottlenecks can be easily identified no matter where they occur.
Phase I: Monitor Local Routed Interfaces
The first step is to monitor any bandwidth constrained links like Internet or WAN links. This should be monitored via the local router interface if at all possible, as it reduces monitoring traffic across the WAN.
At this point, you will be able to track usage on the WAN links, and any received errors as seen from the local router interface.
Phase II: Monitor Remote Routed Interfaces
The next step is to set up monitoring of the remote router interfaces. The purpose of monitoring the far end of the connection is to make sure that interface reception errors are captured. If your local router transmits packets successfully, it does not guarantee that the remote end receives them – it’s valuable to know that the remote end is not discarding mal-formed packets.
Phase III: Monitor all LAN interfaces on Switches
There are many problems that can occur in LANs that many engineers wouldn’t normally suspect:
CAT 3 cabling used by mistake/cabling faults
Links that are mis-configured for 10mbps instead of 100mbps
In addition, monitoring all switched links gives you the facts needed to help justify upgrading to gigabit (or faster) interfaces.
Phase IV: Monitor all interfaces on Servers that support SNMP
Monitoring just the switch interface where servers are connected will disclose any packet reception errors that the server transmitted. To gain confidence that the switch is not transmitting bad packets to the server, the server’s interface should also be monitored.
This usually requires configuring SNMP on the server’s OS and adding it to monitoring.
Note: Some server NICs do not provide accurate SNMP statistics. Different NIC vendors and drivers provide different levels of support of SNMP. NIC drivers should support RFC-1213 statistics to make this useful.
Phase V: Adding Alerts
PathSolutions alerting mechanism is incredibly flexible, as you can set up one wildcard alert for the entire network, or be very specific to a specific interface.