-
Notifications
You must be signed in to change notification settings - Fork 426
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support prometheus #168
base: master
Are you sure you want to change the base?
Support prometheus #168
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is going to break all of our existing monitoring. Is there a way to make this backwards compatible?
+1, maybe its better to wrap promethues metrics to add a prefix. |
also http_ usually implies its related to http endpoints...you should just use kraken_ |
it is difficult to do. i have considered that but it will break the abstract architecture if only add prefix to promethues. i will try. |
you are right. maybe kraken_http_ to identify that it's from kraken and related to http. |
I actually think this change is fine...let's discuss a little more about this |
type PrometheusConfig struct { | ||
ListenAddress string `yaml:"listen_address"` | ||
// HandlerPath if not define use default /metrics. | ||
HandlerPath string `yaml:"handler_path"` |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The yaml config keys are camel cased in the tally package:
Perhaps we should prefer camel case here as well so that it's not a breaking change if we want to swap PrometheusConfig
out with the tally prometheus.Configuration
type in the future.
#161
add 'http_' as prefix of metrics name. for there is a restriction in prometheus that metrics name should start with letter.
and you can set backend to 'promethues' to active prometheus metrics.