Can't see container CPU/Memory usage. #3
Comments
|
What version of docker are you running? Can you take a post logs from the agent and the ui containers. to get the logs you can run
|
|
This is the log for the agent:
For the web ui:
Docker Version
Server: OS Version
|
|
Looks like the UI container's connection to the agents is failing for some
reason. Does the issue resolve itself if you stop and restart the UI
container?
…On Fri, Jun 30, 2017 at 10:43 PM, Shabbir Kagalwala < ***@***.***> wrote:
Reopened #3 <#3>.
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub
<#3 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AMcvTMe-iflk04wY3Gbla5M_eEIIYgp1ks5sJbJsgaJpZM4OLDpP>
.
|
|
I am using a service for the web ui, i did try removing the service, adding it again, adding the nodes/hosts again but still couldn't see the CPU and memory usage. It is strange as I am able to add nodes/hosts, see the containers running on the nodes, can even see the topology (without any connections) but can't see the CPU and memory usage. Are there any additional ports that the web ui uses to communicate with the agent containers except for 9090? |
|
Port 9090 is the only port that needs to be open between the UI service and the agents. The log shows that the websocket is being closed right after it is established. The UI tries to maintain a persistent websocket connection with the agent in order to receive the stats. |
|
So, what could be causing this issue? I have 9090 open in the security groups and can also add nodes.
Is it because I am running the web-ui as a service?
Even when the UI and Agent are on the same host, i cant see the stats for the containers running on the same host.
Thank you for the help so far, I appreciate it!
… On Jul 5, 2017, at 4:55 PM, derhally ***@***.***> wrote:
Port 9090 is the only port that needs to be open between the UI service and the agents. The log shows that the websocket is being closed right after it is established. The UI tries to maintain a persistent websocket connection with the agent in order to receive the stats.
—
You are receiving this because you modified the open/close state.
Reply to this email directly, view it on GitHub, or mute the thread.
|
|
@shabbirkagalwala you bet. Here's another troubleshooting tactic - you can test for connectivity as well as hunt for the statistics you're missing for via the Agent's REST API by pointing your browser at any of these endpoints:
@derhally can point to which of these should have the CPU & memory statistics. |
|
@leecalcote Thank you for the response. I tried all the links as you mentioned and I can see the output. For example:
I guess as mentioned by @derhally the agents not being able to communicate with the web-ui might be the issue. I am going to try tweaking my security groups and Network ACL's to see if anything helps and will get back here to report on what i find. Thank you for all the help so far, truly appreciate it! EDITLooking at the :9090/containers/container-id, this is what i found, it says 404 page not found at the end, could that be an issue?
EDIT 2I opened up all ports and allowed traffic still cant see the stats. I think the ports aren't the issue now, let me know if I need to do anything else to troubleshoot. |
I started the service for swi-ui and containers for the agent using docker run on each node of the swarm. I was also able to add the add nodes using the web-ui but now for some reason, the container stats i.e CPU and Memory just show 0.00% and null values. (see screenshot)
My service create command:
docker service create \ --name swi-ui \ --replicas 1 \ --network webhelpdesk --network proxy \ --label com.df.notify=true --label com.df.distribute=true \ --label com.df.servicePath=/ --label com.df.port=80 \ solarwinds/container-uiDocker run command:
docker run -d --name swi-agent --privileged --net=host --restart always -v /var/run/docker.sock:/host/var/run/docker.sock -v /dev:/host/dev -v /proc:/host/proc:ro solarwinds/container-agentPort 9090 is open and I can see ./agent using this port via the command netstat -tuplen.
Are there any additional ports that are being used to display the container stats? If not, what may be causing the issue here?
The text was updated successfully, but these errors were encountered: