v1

func BlockUntilServerReady ¶

func BlockUntilServerReady(ctx context.Context, addr string, opts ...OpOption) error

func CheckHealthz ¶

func CheckHealthz(ctx context.Context, addr string, opts ...OpOption) error

func DeregisterComponent ¶ added in v0.5.0

func DeregisterComponent(ctx context.Context, addr string, componentName string, opts ...OpOption) error

DeregisterComponent deregisters the component from the server, by using the component name. It fails if the component has not been registered yet or is not deregisterable.

func GetComponents ¶

func GetComponents(ctx context.Context, addr string, opts ...OpOption) ([]string, error)

func GetEvents ¶

func GetEvents(ctx context.Context, addr string, opts ...OpOption) (v1.GPUdComponentEvents, error)

func GetHealthStates ¶ added in v0.5.0

func GetHealthStates(ctx context.Context, addr string, opts ...OpOption) (v1.GPUdComponentHealthStates, error)

GetHealthStates returns health states from the server for the given components.

Example:

baseURL := "https://localhost:15132"
for _, componentName := range []string{"disk", "accelerator-nvidia-info"} {
	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
	defer cancel()
	states, err := GetHealthStates(ctx, baseURL, WithComponent(componentName))
	if err != nil {
		if errdefs.IsNotFound(err) {
			log.Logger.Warnw("component not found", "component", componentName)
			return
		}

		log.Logger.Error("error fetching component info", "error", err)
		return
	}

	for _, ss := range states {
		for _, s := range ss.States {
			log.Logger.Infof("state: %q, health: %s\n", s.Name, s.Health)
		}
	}
}

func GetInfo ¶

func GetInfo(ctx context.Context, addr string, opts ...OpOption) (v1.GPUdComponentInfos, error)

GetInfo returns component information from the server.

Example:

baseURL := "https://localhost:15132"
componentName := "" // Leave empty to query all components

ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
defer cancel()
info, err := GetInfo(ctx, baseURL, WithComponent(componentName))
if err != nil {
	fmt.Println("Error fetching component info:", err)
	return
}

fmt.Println("Component Information:")
for _, i := range info {
	fmt.Printf("Component: %s\n", i.Component)
	for _, event := range i.Info.Events {
		fmt.Printf("  Event: %s - %s\n", event.Name, event.Message)
	}
	for _, metric := range i.Info.Metrics {
		fmt.Printf("  Metric: %s (labels: %q) - Value: %f\n", metric.Name, metric.Labels, metric.Value)
	}
	for _, state := range i.Info.States {
		fmt.Printf("  State: %s - Health: %s\n", state.Name, state.Health)
	}
}

func GetMachineInfo ¶ added in v0.5.0

func GetMachineInfo(ctx context.Context, addr string, opts ...OpOption) (*apiv1.MachineInfo, error)

func GetMetrics ¶

func GetMetrics(ctx context.Context, addr string, opts ...OpOption) (v1.GPUdComponentMetrics, error)

func GetPackageStatus ¶ added in v0.4.3

func GetPackageStatus(ctx context.Context, url string, opts ...OpOption) ([]packages.PackageStatus, error)

GetPackageStatus fetches the GPUd package status from the GPUd admin API.

func GetPluginSpecs ¶ added in v0.5.0

func GetPluginSpecs(ctx context.Context, addr string, opts ...OpOption) (pkgcustomplugins.Specs, error)

GetPluginSpecs returns the custom plugins registered in the server.

func ReadComponents ¶

func ReadComponents(rd io.Reader, opts ...OpOption) ([]string, error)

func ReadEvents ¶

func ReadEvents(rd io.Reader, opts ...OpOption) (v1.GPUdComponentEvents, error)

func ReadHealthStates ¶ added in v0.5.0

func ReadHealthStates(rd io.Reader, opts ...OpOption) (v1.GPUdComponentHealthStates, error)

func ReadInfo ¶

func ReadInfo(rd io.Reader, opts ...OpOption) (v1.GPUdComponentInfos, error)

func ReadMetrics ¶

func ReadMetrics(rd io.Reader, opts ...OpOption) (v1.GPUdComponentMetrics, error)

func ReadPluginSpecs ¶ added in v0.5.0

func ReadPluginSpecs(rd io.Reader, opts ...OpOption) (pkgcustomplugins.Specs, error)

ReadPluginSpecs reads the custom plugin specs from the server.

func SetHealthyComponents ¶ added in v0.8.0

func SetHealthyComponents(ctx context.Context, addr string, components []string, opts ...OpOption) ([]string, error)

SetHealthyComponents sets specified components to healthy state

func TriggerComponent ¶ added in v0.5.0

func TriggerComponent(ctx context.Context, addr string, componentName string, opts ...OpOption) (v1.GPUdComponentHealthStates, error)

TriggerComponent manually triggers a component check.

func TriggerComponentCheckByTag ¶ added in v0.5.0

func TriggerComponentCheckByTag(ctx context.Context, addr string, tagName string, opts ...OpOption) error

TriggerComponentCheckByTag triggers all components that have the specified tag