Monitoring and alerting manual

The SNMP module consists of 3 different pollers:

A. An ICMP ping poller that will simply ping all the configured devices and provide 4 parameters:

• Packet loss for each per-minute ping.
• Packet roundtrip delay for each per-minute ping.
• Average packet loss during an hour.
• Average roundtrip delay during an hour.

B. A TCP port probe that will send packets to the port specified and provide the same parameters as above.

C. The very SNMP poller that will query each device for the OIDs configured for it, and will show OIDs values as is in the web interface, or calculate various other parameters using formulas defined by the user.

Some of the parameters per-configured in 5gVision are:

• CPU average over 1 min, %.
• CPU average over each hour, %.
• RAM used (w/o cache), %.
• HDD root used, %.
• In/Out KBytes per minute/second
• In/Out MBytes per hour

The 5gVision SNMP module delivers very flexible configuration options and can be configured to only poll certain devices for certain parameters via group assignments. Raw OID values can be transformed to parameters via formulas.

The core principle of the 5gVision SNMP module is that there are two types of groups that can be arbitrary linked to each other: Device groups, and Parameter groups. A device may belong to several groups or parameters. You may also link individual devices to individual parameters or groups, if needed. This setup results in a very flexible way of determining which devices will have which parameters polled.

SNMP statistics, once collected, will be no different from the statistics extracted from the VoIP switch CDRs, so the modules discussed in this Manual: Chart, Table, Report, Alerts will show SNMP statistics in much the same way.

Devices can be any piece of equipment having an IP or a domain name. If the device is running SNMP daemon, or sending traps, it can be configured for collection of SNMP statistics, if not - you can only monitor ping packet loss and response time. The "ping" method can actually be not only the ICMP ping, but also probing a specific port open at a device. This is convenient when a device has ping responses blocked.

These groups are required to logically unite similar devices together, so that you do not need to link each individual device to each individual parameter that will be collected.

OIDs table is simply a collection of all OIDs that will be polled at least for one device. Each device group will be polled only for the specific OIDs determined for each parameter, that is in turn linked to the device group via the parameter group. If you wish to use OIDs in formulas to calculate parameter values - you need to assign a variable name to each OID.

Parameters can take the exact values from the SNMP response on each OID, or can be obtained via the following transformation methods:

• By creating a formula using existing OID values with variable names assigned to them (in_bytes/1024, hdd_root_used/hdd_root_total*100, etc.).
• By calculating the delta value between the current and the previous OID values (since polling is done every minute, the "previous" values are the values collected one minute before).
• By calculating per-hour parameter values from collected per-minute values using the following methods: total, average, minimum, maximum for the hour.

If Parameter OID name field is not empty, the parameter value will be taken as is from the SNMP response, and any entries in the Parameter formula field will be ignored. Otherwise, the Parameter formula will determine how the parameter will be calculated from one, or several variable names associated with the OIDs.

You may use the most common arithmetic operations in formulas, like: +, -, *, /. Support for more complex expressions, exponential functions, etc. is not guaranteed, but may be tested and added upon request.

These groups are required to logically unite similar parameters together, so that you do not need to link each individual device to each individual parameter that will be collected.

The links will determine which device groups (or an individual device) will be polled for which specific parameters in a parameter group (or for just one individual parameter). Links table works on full-mesh principles, if you specify all 4 entries: a device group, a device, a parameter group, and a parameter in one row, all 4 of them will be linked to each other.