You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Plugins can independently expose Prometheus or other types of metrics that can be scraped or sent completely separately from GatewayD, however, it would be nice to have a feature to aggregate all Prometheus metrics in GatewayD, and then expose them on /metrics all at once.
This is my proposal:
Each plugin creates and exposes their own Prometheus metrics on the /metrics HTTP endpoint on a Unix/TCP socket.
The configuration will be passed to GatewayD as part of their configs.
While loading the plugin on startup, GatewayD will register the Unix socket's address or TCP port.
The merged and relabeled metrics from plugins, along with the ones from GatewayD, will be collected and exposed on the HTTP /metrics endpoints and served on TCP in GatewayD.
Plugins can independently expose Prometheus or other types of metrics that can be scraped or sent completely separately from GatewayD, however, it would be nice to have a feature to aggregate all Prometheus metrics in GatewayD, and then expose them on
/metrics
all at once.This is my proposal:
Each plugin creates and exposes their own Prometheus metrics on the
/metrics
HTTP endpoint on a Unix/TCP socket.The configuration will be passed to GatewayD as part of their
configs
.While loading the plugin on startup, GatewayD will register the Unix socket's address or TCP port.
The merged and relabeled metrics from plugins, along with the ones from GatewayD, will be collected and exposed on the HTTP
/metrics
endpoints and served on TCP in GatewayD.Metrics export gatewayd-plugin-sdk#2
The text was updated successfully, but these errors were encountered: