After not using a broker, the endpoints are unable to connect to the server.

[Chilla]

Dear community,

I tried to make some agents connect directly to the server without using a broker. 

However, it's not able not connect to server.

I referred to the practices of others, and tried executing "cytool reconnect force", but still couldn't establish a connection. Because some endpoints can be successfully connected, it can be ruled out that the firewall is blocking the connection.

After checking the traps.log,

I noticed that the logs do not mention FQDN. Is this normal?

Furthermore, the broker has been confirmed to be offline, but it appears that it is still attempting to connect to the broker and showing a "connect timed out" error.

Upon checking the "All endpoints" section, it seems that the proxy field still contains the broker IP. Could this be related to the issue? But why is it that some endpoints can still connect directly to the server, even though the proxy field still contains a record?

thx for any help.

[neelrohit]

Hi @Chilla ,

Since this a public forum, I would recommend removing this post or sensitive information(logs) from the post. Also, I would recommend opening a support case with our support team to investigate the same. 

Also you might want to check with your network team if they have allowed cortex urls to facilitate connection without broker vm. Generally air gapped systems do not allow firewall rules to facilitate direct connections from  the endpoints to the cloud.

Also, even if the broker VM is down and the proxy configuration is set, the agent will first attempt communication to the cloud via broker 2 times. Upon 2 failures, it will attempt direct connection to the cloud. However, this setting can also be overidden to prevent direct communication and you might want to check if the agent settings policy configuration has direct server access disabled or not.

Hope this helps!