Hi Janaka, Jerzy, Sander and everyone,

There is something unclear for me to understand the setup so I just wanna 
ask:

- Do we need Prometheus running in order to do hpa with custom metric 
(queries per second) ? OR cAdvisor and heapster are enough ?
- By binding the configMap to a pod and adding the annotation to HPA 
manifest, we will be good to have it autoscaled? 

-Tuna

On Thursday, June 16, 2016 at 1:59:28 PM UTC+2, Sander Ploegsma wrote:
>
> EDIT: never mind, my endpoint was being served over HTTPS, which made it 
> insecure as the certificate wasn't valid for 'localhost'. When using HTTP 
> everything works as expected. Marking as resolved.
>
> Hi, Janaka,
>
> Could you show how you got it working exactly? I tried setting everything 
> up according to the documentation proposed in the PR, but it's a bit 
> unclear how the endpoint should be structured.  
>
> My steps:
>
>    - Enabled custom metrics on the cluster
>    - Created an endpoint on my application that returns the number of 
>    open connections
>    - Created this exact ConfigMap: 
>    apiVersion: v1
>    kind: ConfigMap
>    metadata:
>      name: cm-config
>    data:
>      # Use Prometheus structure, see 
>    https://github.com/google/cadvisor/blob/master/docs/application_metrics.md
>      definition.json: "{\"endpoint\" : \"http://localhost:8443/metrics\";, 
>    {\"metrics_config\" : [{\"name\" : \"connections\", \"metric_type\" : 
>    \"gauge\", \"data_type\" : \"int\", \"polling_frequency\" : 10, \"units\" 
> : 
>    \"number of active connections\"}]}}"
>    
>    - Mapped the hostPort and containerPort in the RC
>    - Mounted the ConfigMap in my ReplicationController to 
>    /etc/custom-metrics
>    - Added this exact annotation to my HPA: 
>    alpha/target.custom-metrics.podautoscaler.kubernetes.io: 
> '{"items":[{"name":"connections", 
>    "value": "1000"}]}'
>    
> However, the HPA is unable to get my custom metrics. I also couldn't find 
> any way to debug the setup, all I see is 'FailedGetCustomMetrics metrics 
> obtained for 0/1 of pods' when I check the HPA events.
>
> Cheers, 
>
> Sander
>
> On Wednesday, May 4, 2016 at 6:17:50 PM UTC+2, Janaka Bandara wrote:
>>
>> Hi Jerzy,
>>
>> Thanks!
>> I tried the approach in the PR and it worked fine.
>>
>>
>> (I found the same links a few days ago from a different source, but could 
>> not update this thread on time.)
>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"Kubernetes user discussion and Q&A" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to kubernetes-users+unsubscr...@googlegroups.com.
To post to this group, send email to kubernetes-users@googlegroups.com.
Visit this group at https://groups.google.com/group/kubernetes-users.
For more options, visit https://groups.google.com/d/optout.
  • [kubernete... Tuna
    • [kube... Tuna
      • [... 'Jerzy Szczepkowski' via Kubernetes user discussion and Q&A

Reply via email to