Node successfully registered, but does not send metrics

  1. Verify first that your question wasn’t asked before on the forum.
    I believe so, as all the results I found are having issues registering the node, while I can do that, but THEN netdata cloud is not getting metrics.

  2. Briefly explain what is the problem you are having
    I am able to register a node, but not get any metrics about it in Netdata Cloud and it shows offline despite the fact that I am actively SSH’d in.

  3. Try to provide as much information about your environment as possible. (e.g OS, container, etc.)
    Ubuntu 18.04
    Installed via APT, updated…like last week or something to the new Netdata version.

  4. Select the appropriate tags to specify the components that are affected
    I hope my tags are correct, I have no idea whether the problem is with the agent or with the cloud or both.

  5. Provide a clear and concise description of what you expected to happen.
    When I received this message when successfully registering a node, I assumed my node can reach your cloud and should be able to send metrics in and show as online. Instead, it is showing completely offline, despite the fact that it IS online, I can SSH the server, and it can register or re-register the server as a node successfully to my Netdata Cloud. Re-registering does not fix the issue.

Netdata user: netdata
Connection attempt 1 successful
Node was successfully claimed.

Hi @Mbrantley!
I think the fact that your installed Netdata via APT may cause the problem since this is not supported by us at the time.
The best way to troubleshoot is to try and uninstall Netdata and then try to install it from scratch by using the kickstart script, as it appears in your Cloud app when you click “Connect Nodes”. This way you’ll be able to both install the agent and directly connect it to the Cloud.

Please let me know if this workaround helps and also share with us the Netdata error logs in case this fails.

" To see the full list of approved methods for each operating system/version we support, see our distribution matrix."

If this is not supported, why is your team putting effort into creating these repositories and then documenting how to install Netdata with it? I understand that documentation isn’t always up to date and correct, but it very strongly implies that there is no real downside to using .deb other than you cannot use Xenstats & AWS Kinesis as indicated on the distro matrix page I linked.

If I MUST use kickstart.sh to get any sort of support, I can accept that this one node just will never work. Sorry to waste your time but I’m not interested in installing software outside of my package manager as a general rule. I would lightly suggest putting warnings on the pages I linked that despite it being in your documentation, that it is not recommended and that no support is available if you follow those instructions to install netdata. Even just yellow “Yield” icons or something to scare people away from using them would seem appropriate and probably lead to less of your users installing Netdata in ways you actually don’t want us doing.

You can probably also delete my other support thread as that server also was installed via APT using your packagecloud repo and so it’s probably useless to you:

Hi @Mbrantley,

Could you please provide the netdata agent version and some logs so that we can troubleshoot what the problem is?

If it’s easier for you, you can also contact us on netdata discord .

Hi, @Mbrantley first of all sorry to hear you are having issues.

I will try and install VM with Ubuntu 18.04 and take a look myself but if you could provide the output of the following commands it would be awesome (perhaps you will be quicker than me installing the VM):

  1. netdata -v to get the version of netdata you use
  2. netdata -W buildinfo to get information of how netdata is built by package mantainer of your distro
  3. netdatacli aclk-state - this one needs to be used while netdata daemon/service is already running and might not be available if your netdata version is too old.

Thanks a lot

Installing netdata in fresh install of Ubuntu 18.04 installs netdata v.1.9.0, current stable version of netdata is 1.32.1 so the version in repositories is very very old and can’t connect to the current cloud.

Attempting to use cloud provided claim command (to register noted with cloud space) yields to error this in my case (which is valid):

bash <(curl -Ss https://my-netdata.io/kickstart.sh) --claim-token XXXXXXXXXXXX-XXXXXXXXXXXXX  --claim-rooms XXXXXX-4ab0-43dc-9754-faf517e6d506 --claim-url https://app.netdata.cloud
 --- Found existing install of Netdata under: / --- 
 --- Existing install appears to be handled manually or through the system package manager. --- 
 --- Attempting to claim agent to https://staging.netdata.cloud --- 
sudo: /usr/sbin/netdata-claim.sh: command not found
 FAILED  

I’m not installing it from the Ubuntu default repositories. I followed these instructions here to use the packagecloud repositories that your team puts together:
(I assume someone on your team puts these repos together because I learned about these repositories by reading your documentation which I have linked right here)

The exact path of the repository I am pulling from is:

You can see all the different OS packages in this repo that are version 1.32.1, which I updated to this morning, and yesterday I was on v1.32.0. This is exactly how I have installed and exactly what version I am on, the current stable release, always within 24 hours of a release. Sorry that you spun up a VM just to test all that.

This morning, the node in question is working just fine, so, I’m completely at a loss as to what I was seeing yesterday. I’m assuming it must have been an issue in v1.32.0 that no longer occurs in v1.32.1.

I’m probably being a huge pain and I’m very sorry, I’m just at a loss that no one seems to know what I’m talking about. I would never install from the default ubuntu repositories because your documentation on using apt/rpm says, explicitly, to use the packagecloud repositories which made sense to me, as it has the current stable release.

Everything I’m saying I learned by comprehensively reading your installation documentation and ensuring I was able to install the current stable release channel through apt via the packagecloud repositories so that I can have granular control over the exact hour that I install said updates to ensure that the services that my Nodes deliver are never affected whatsoever by any update occurring out of my control consuming bandwidth or system resources when it matters.

I rejoined the discord, I left it after I thought I was ineligible for any support. My username is “Ralt” if you want to reach out for any reason, I know we’re in different time zones but I’m very often up quite late.

Hi @Mbrantley , I am very sorry for the misunderstanding and the fact I assumed that you are using a completely different method of installing Netdata.

We are definitely happy to assist all our users and feel free to reach out anytime you have a question and may have questions about the setup and our features. Moreover, all our users are eligible for support both here and on Discord, so please let us know about your challenges and concerns with monitoring and Netdata.

Regarding your node that seemed to be unreachable, I’m happy to hear that you are not encountering the issue anymore, and we are definitely at your disposal for further questions.