metricdispatch

func Init ¶

func Init(rawConfig json.RawMessage) error

func LoadAverages ¶

func LoadAverages(
	job *schema.Job,
	metrics []string,
	data [][]schema.Float,
	ctx context.Context,
) error

LoadAverages computes average values for the specified metrics across all nodes of a job. For running jobs, it loads statistics from the metric store. For completed jobs, it uses the pre-calculated averages from the job archive. The results are appended to the data slice.

func LoadData ¶

func LoadData(job *schema.Job,
	metrics []string,
	scopes []schema.MetricScope,
	ctx context.Context,
	resolution int,
) (schema.JobData, error)

LoadData retrieves metric data for a job from the appropriate backend (memory store for running jobs, archive for completed jobs) and applies caching, resampling, and statistics generation as needed.

For running jobs or when archive is disabled, data is fetched from the metric store. For completed archived jobs, data is loaded from the job archive and resampled if needed.

Parameters:

• job: The job for which to load metric data
• metrics: List of metric names to load (nil loads all metrics for the cluster)
• scopes: Metric scopes to include (nil defaults to node scope)
• ctx: Context for cancellation and timeouts
• resolution: Target number of data points for resampling (only applies to archived data)

Returns the loaded job data and any error encountered. For partial errors (some metrics failed), the function returns the successfully loaded data with a warning logged.

func LoadJobStats ¶

func LoadJobStats(
	job *schema.Job,
	metrics []string,
	ctx context.Context,
) (map[string]schema.MetricStatistics, error)

LoadJobStats retrieves aggregated statistics (min/avg/max) for each requested metric across all job nodes. For running jobs, statistics are computed from the metric store. For completed jobs, pre-calculated statistics are loaded from the job archive.

func LoadNodeData ¶

func LoadNodeData(
	cluster string,
	metrics, nodes []string,
	scopes []schema.MetricScope,
	from, to time.Time,
	ctx context.Context,
) (map[string]map[string][]*schema.JobMetric, error)

LoadNodeData retrieves metric data for specific nodes in a cluster within a time range. This is used for node monitoring views and system status pages. Data is always fetched from the metric store (not the archive) since it's for current/recent node status monitoring.

Returns a nested map structure: node -> metric -> scoped data. FIXME: Add support for subcluster specific cc-metric-stores

func LoadNodeListData ¶

func LoadNodeListData(
	cluster, subCluster string,
	nodes []string,
	metrics []string,
	scopes []schema.MetricScope,
	resolution int,
	from, to time.Time,
	ctx context.Context,
) (map[string]schema.JobData, error)

LoadNodeListData retrieves time-series metric data for multiple nodes within a time range, with optional resampling and automatic statistics generation for large datasets. This is used for comparing multiple nodes or displaying node status over time.

Returns a map of node names to their job-like metric data structures.

func LoadScopedJobStats ¶

func LoadScopedJobStats(
	job *schema.Job,
	metrics []string,
	scopes []schema.MetricScope,
	ctx context.Context,
) (schema.ScopedJobStats, error)

LoadScopedJobStats retrieves job statistics organized by metric scope (node, socket, core, accelerator). For running jobs, statistics are computed from the metric store. For completed jobs, pre-calculated statistics are loaded from the job archive.