Creating a TCP monitor

The NS1 monitoring solution allows you to create monitors that track the health and performance of a specified device or endpoint. users can create any of the following monitor types:

• PING
• HTTP/HTTPS
• TCP

Server/device TCP monitoring tracks the availability and response time of a TCP server from a local network—for example, if a TCP server is deployed within an intranet and is not accessible outside of it.

This article explains the process for setting up a TCP monitor and connecting it to answer metadata.

Instructions

Step 1: Create a TCP monitor.

1. In the NS1 portal, navigate to the Monitors tab.

2. Click Add a Monitor.

3. Select TCP from the drop-down menu in the upper-left corner.
   
   
   
   

4. Under Configuration,

   • Enter a Name for the monitor.

   • Optionally, toggle the button next to Monitor Paused in order to active or deactivate the monitor. (Default is disabled, meaning the monitor is active.)

   • Toggle the button next to Notifications On to enable or disable notifications related to this monitor. (Default is enabled, meaning notifications are active.)

   • Select the desired monitoring regions. These are the locations from which the monitoring will be executed.

   • Under Policy, select one of the following from the drop-down menu:

     • Quorum: The host is marked “down” if tests conducted from a majority of the monitoring regions do not pass the “up” conditions.’

     • All: The host is marked “down” if tests conducted from all of the monitoring regions do not pass the “up” conditions.

     • One: The host is marked “down” if tests conducted from a single monitoring region do not pass the “up” conditions.

   • Under Frequency, enter the amount of time (in seconds) between each monitoring test conducted in each region.

   • Under Up Conditions, click Add Condition to define the conditions the hostname must meet in order to be considered “up.”

     First, choose the metric to validate during the test:

     - Timeout: The time (in milliseconds) for the connection between the server and client to open.
     - Output: Output received from any connection, if any.

     Next, choose the comparison operator (=, <, >, etc), and the value to compare. You can add multiple “up” conditions, just note that all conditions must be met in order for a test run to consider the host “up.”

   • Optionally, check the box next to Rapid recheck to automate a second verification test before changing the status of a host. Checking this box can help prevent false positives.

5. Under TCP Settings,

   • The response timeout is the time (in milliseconds) after connecting to wait for output.

   • The string to send is the string sent to hosts upon connecting. Note that string escapes (ex. /n) are allowed.

   • Enter a TCP port. If you are monitoring an HTTP server, set the TCP port to 80. For an HTTPS server, set the TCP port to 443.

   • Enable or disable the option to connect with SSL.

   • Enter the IP address (IPv4 or IPv6) or fully-qualified domain name of the device you would like to monitor.

   • Optionally, check the both next to Connect over IPv6. If checked, verify that you entered either a fully qualified domain name or IPv6 address.

     NOTEIf you select this option, the monitor will connect exclusively over IPv6.

   • Optionally, enter a value under Connect timeout to indicate the time (in milliseconds) before a host should be marked as “failed” or “down.”

6. Once complete, click Save Changes. The new TCP monitor appears in the list.

Step 2: Connect the monitor to an answer using the NS1 Filter Chain.

To complete the setup process, you must connect the monitor to an answer in order to automatically change the “up” metadata to reflect the status of the monitor. That is, when the tested host meets the up conditions, it will change the metadata to up=true. If the tested host does not meet the up conditions, the metadata is automatically changed to up=false.

1. Under Zones, navigate to the record upon which you want to attach the monitor. Then, click into the record to view associated answers.
   
   
   
   
2. Click Create Filter Chain.
   
   
   
   
3. Select or drag-and-drop the Up filter option (and any other desired filters) to the list of active filters.
   
   
   
   
4. Click Save Filter Chain. Now, you can see the Filter Chain configuration to the left of the answer list.
   
   
   
   
5. On the left-hand sidebar showing the Filter Chain configuration, click the UP filter to expand that box and reveal filter-specific metadata options.
   
   
   
   
6. Click the up: unset metadata label beneath the answer to view the Answer Metadata configuration screen.
   
   
   
   
   
   
7. Click the “feed” icon next to the up/down filter setting to reveal a list of available data feeds—including the monitor you created earlier.
   
   
   
   
8. Select the TCP monitor you created earlier, and click Ok.
   
   
   
   
9. Click Save Record to save your changes.
   
   

Now, when the TCP monitor detects a failed endpoint, it automatically adjusts the answer metadata to indicate that the endpoint is “down.” As a result, DNS traffic is steered to better-performing endpoints.