Grafana

Grafana is a dashboard tool that can query data stored in a database and display data in real time in various display formats such as graphs, numbers, tables, and text. Various databases are supported, such as InfluxDB,Prometheus,MySQL, etc., and multiple data sources can be mixed in the same graph.

It is also rich in functionality, and representative ones include an alert function that notifies externally of data information when a threshold is exceeded, and a function that outputs dashboard visualization data to PDF.

Figure 2: Grafana dashboard visualization example

Telegraph

Telegraf is an agent that periodically collects resource data of the target device and stores it in a database such as InfluxDB. It supports a wide range of protocols such as SNMP and gNMI. Since it is highly extensible, it is also possible to add original functions.

InfluxDB

InfluxDB is a time-series database that is useful when dealing with time-linked data. It is one of the representative data sources in Grafana, and the database can be controlled by command line or GUI.

SONiC

The settings on the SONiC side are very simple as shown below.
In the configuration example, the command is executed assuming SNMPv2c, but the required settings differ depending on the version used.

admin@sonic:~$ sudo config interface ip add eth0 192.168.0.60/24 192.168.0.254 /* マネジメントポートへの IP アドレス設定 */ admin@sonic:~$ sudo config snmp community add macnica ro /* SNMP コミュニティを作成 */

OSS various tools

Telegraf requires the following settings changes after installing the tool. Other tools work with the default settings.
・Enabling SNMP function
- MIB specification for SNMP monitoring
・Polling interval settings

Visualization by traffic

In this example, the following traffic data received by the interface is visualized on the dashboard.
・Total number of received traffic (IF Traffic)
・Total number of discarded traffic (IF Discards)

As mentioned above, Grafana collects data from the database at regular intervals, so the data is reflected in the dashboard in real time.
The graph below shows the communication status of the Ethernet124 interface.

Figure 3: Ethernet124 communication status

It visualizes the 11-minute communication status received on the interface from 09:39:00 to 09:50:00, and the Y-axis shows the number of received traffic octets and the number of discards, respectively. Checking the IF Traffic graph, a certain amount of traffic was received from 09:40:00 and then continued until 09:50:00. You can see at a glance that no waste has occurred.

Variable filtering function

With Grafana, you can switch the display contents of multiple dashboards by filtering with registered variables. As an example, use the device interface name as a variable to toggle the dashboard content. Earlier I illustrated the data of the traffic received on Ethernet124, but the following is the data of the traffic received on Ethernet120.

Figure 4: Ethernet120 communication status

It is a variable in which items such as loop and CPU displayed above the figure are registered, and it is possible to switch variables with a drop-down formula. The part surrounded by a red frame is the variable used for switching this time. Comparing the two diagrams, you can see that the dashboard switch was successful as the data displayed is different. If the data to be displayed in this way overlaps in units such as interfaces, it is possible to easily change the displayed data by switching variables without creating a new dashboard.

Visualization that aggregates multiple traffic

Grafana The data for multiple traffic 1 It is also possible to consolidate and visualize in one panel.
The graph below shows the TCP connection status for one day.

・Green line: Monitored machine is active (From own aircraft to other aircraft) Open a connection to
・Yellow line: Monitored aircraft is passive (From another machine to your own machine) Open a connection to
・Blue line: Connection establishment failure
・Orange line: Connection reset

Figure 5: TCP communication status

The Y-axis represents the total amount of each TCP traffic, and if you look at the graph, you can see that the frequency of opening and failing all connections is slightly higher around 12:00. By consolidating highly related data into one panel in this way, data analysis can be performed efficiently.

This concludes the explanation of the network visualization environment for white Box switches using OSS. With ecSONiC, which was introduced in Part 5, it is possible to build an environment using the same procedure as the OSS version of SONiC. In future articles, we plan to introduce the network visualization environment combined with OcNOS provided by IP Infusion and know-how on creating a Grafana dashboard.

These people are using the remote verification service.

Here are some testimonials from people who have actually used the service.

 

Furukawa Network Solution Co., Ltd.

“Recently, the number of remote service environments is increasing, but I was impressed with the ease of access to the evaluation equipment.

The materials you provided were easy to understand, and we were able to proceed smoothly with the intended verification. ”

Related information

How to visualize SNMP data using OSS ~ Grafana x OcN...

OS～

~

Network visualization dashboard using Grafana for beginners...

code creation procedure

next-generation model...

Interface~

Network visualization of white Box switches using commercial tools ~SolarWinds edition~

～Grafana x InfluxDB Edition～