Documentation
¶
Index ¶
- func NewExemplarQueryable(upstream storage.ExemplarQueryable, bypassWithSingleQuerier bool, ...) storage.ExemplarQueryable
- func NewMergeExemplarQueryable(idLabelName string, upstream storage.ExemplarQueryable, ...) storage.ExemplarQueryable
- func NewMergeQueryable(idLabelName string, callbacks MergeQueryableCallbacks, ...) storage.Queryable
- func NewMetadataSupplier(next querier.MetadataSupplier, maxConcurrency int, logger log.Logger) querier.MetadataSupplier
- func NewQueryable(upstream storage.Queryable, bypassWithSingleID bool, maxConcurrency int, ...) storage.Queryable
- type Config
- type MergeQuerierUpstream
- type MergeQueryableCallbacks
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func NewExemplarQueryable ¶
func NewExemplarQueryable(upstream storage.ExemplarQueryable, bypassWithSingleQuerier bool, maxConcurrency int, reg prometheus.Registerer, logger log.Logger) storage.ExemplarQueryable
NewExemplarQueryable returns an exemplar queryable that makes requests for all tenant IDs that are part of the request and aggregates the results from each tenant's ExemplarQuerier.
Results contain the series label __tenant_id__ to identify which tenant the exemplar is from.
By setting bypassWithSingleQuerier to true, tenant federation logic gets bypassed if the request is only for a single tenant. The requests will also not contain the pseudo series label __tenant_id__ in this case.
func NewMergeExemplarQueryable ¶
func NewMergeExemplarQueryable(idLabelName string, upstream storage.ExemplarQueryable, bypassWithSingleQuerier bool, maxConcurrency int, reg prometheus.Registerer, logger log.Logger) storage.ExemplarQueryable
NewMergeExemplarQueryable returns an exemplar queryable that makes requests for all tenant IDs that are part of the request and aggregates the results from each tenant's ExemplarQuerier.
Results contain the series label `idLabelName` to identify which tenant the exemplar is from.
By setting bypassWithSingleQuerier to true, tenant federation logic gets bypassed if the request is only for a single tenant. The requests will also not contain the pseudo series label `idLabelName` in this case.
func NewMergeQueryable ¶
func NewMergeQueryable(idLabelName string, callbacks MergeQueryableCallbacks, resolver tenant.Resolver, bypassWithSingleID bool, maxConcurrency int, reg prometheus.Registerer, logger log.Logger) storage.Queryable
NewMergeQueryable returns a queryable that merges results for all involved federation IDs. The underlying querier is returned by a callback in MergeQueryableCallbacks.
By setting bypassWithSingleID to true the mergeQuerier gets bypassed, and results for requests with a single ID will not contain the ID label. This allows for a smoother transition, when enabling tenant federation in a cluster.
Each result contains a label `idLabelName` to identify the federation ID it originally resulted from. If the label `idLabelName` already exists, its value is overwritten and the previous value is exposed through a new label prefixed with "original_". This behaviour is not implemented recursively.
func NewMetadataSupplier ¶
func NewMetadataSupplier(next querier.MetadataSupplier, maxConcurrency int, logger log.Logger) querier.MetadataSupplier
NewMetadataSupplier returns a querier.MetadataSupplier that returns metric metadata for all tenant IDs that are part of the request and merges the results.
No deduplication of metadata is done before being returned.
func NewQueryable ¶
func NewQueryable(upstream storage.Queryable, bypassWithSingleID bool, maxConcurrency int, reg prometheus.Registerer, logger log.Logger) storage.Queryable
NewQueryable returns a queryable that iterates through all the tenant IDs that are part of the request and aggregates the query results for tenant. By setting bypassWithSingleID to true the mergeQuerier gets bypassed and results for requests with a single ID will not contain the "__tenant_id__" label. This allows for a smoother transition, when enabling tenant federation in a cluster. The result contains a label "__tenant_id__" to identify the tenant ID that it originally resulted from. If the label "__tenant_id__" already exists, its value is overwritten by the tenant ID and the previous value is exposed through a new label prefixed with "original_". This behaviour is not implemented recursively.
Types ¶
type Config ¶
type Config struct {
// Enabled switches on support for multi tenant query federation
Enabled bool `yaml:"enabled"`
MaxConcurrent int `yaml:"max_concurrent" category:"experimental"`
MaxTenants int `yaml:"max_tenants" category:"experimental"`
}
func (*Config) RegisterFlags ¶
type MergeQuerierUpstream ¶
type MergeQuerierUpstream interface {
Select(ctx context.Context, id string, sortSeries bool, hints *storage.SelectHints, matchers ...*labels.Matcher) storage.SeriesSet
LabelValues(ctx context.Context, id string, name string, matchers ...*labels.Matcher) ([]string, annotations.Annotations, error)
LabelNames(ctx context.Context, id string, matchers ...*labels.Matcher) ([]string, annotations.Annotations, error)
Close() error
}
MergeQuerierUpstream mirrors storage.Querier, except every query method also takes a federation ID.
type MergeQueryableCallbacks ¶
type MergeQueryableCallbacks struct {
// Querier returns a MergeQuerierUpstream implementation for mint and maxt.
Querier func(mint, maxt int64) (MergeQuerierUpstream, error)
}
MergeQueryableCallbacks contains callbacks to NewMergeQueryable, for customizing its behaviour.
Source Files