remove coreos-operator dependancy #19
Conversation
|
@krasi-georgiev Wasn't sure if you meant me to try it out yet, or still work in progress - but gave it a quick try. Looks good - got the results in prom, although for some reason the jaeger UI just spins, nothing to do with your PR. Before doing the final PR, I think it would be good to change the label "frontend" to something more meaningful - I think that was just a name prom-operator used in their example. It just needs to be something metric related I guess. |
|
Great I will continue with openshift and will replace the label. |
small bug with the label names
|
all done , I tested both just to make sure. The jaeger UI worked once and than it gets stuck at the spinning circle. |
Kubernetes.md
Outdated
| ``` | ||
|
|
||
| Add configuration to locate service monitors based on label "team: frontend": | ||
| Add configuration to locate service monitors based on label "scrape: true": |
There was a problem hiding this comment.
nit: "locate services to be monitored based..."
OpenShift.md
Outdated
| To obtain metrics from the deployed services, we will use the | ||
| [coreos prometheus operator](https://coreos.com/operators/prometheus/docs/latest/user-guides/getting-started.html) | ||
| project. To install, use the following command: | ||
| Add configuration to locate service monitors based on label "scrape: true": |
|
|
||
| NOTE: The alpha version is currently required, as the previous stable version does not include Kubernetes RBAC authorizaion | ||
| support. | ||
|
|
||
| When fully started, then launch the dashboard using the link displayed following `The server is accessible via web console at: | ||
| `. Log in using the credentials `developer/developer` and navigate to the `New Project` overview page. | ||
|
|
There was a problem hiding this comment.
The next line contains "Kubernetes", could you change this to "OpenShift" - looks like I missed that (copy/paste) :)
|
@krasi-georgiev Great job - just a few minor comments. However the main issue is that the metrics scraped by prometheus, vs prom-operator, are lacking labels for "service", "pod" and "namespace". There is an 'instance' label - but that doesn't seem to be specific to the individual pod - so if we have multiple instances of a particular service, there is no way to determine which instance actually executed the request, Even more important is 'service' (and 'namespace' to a less extent). Any ideas of how these labels could be added in, without the application/service itself including them into the metrics? |
…the prometheus official examples.
TODO: