actions.odigos.io/v1alpha1

Represents the observations of a addclusterinfo’s current state. Known .status.conditions.type are: “Available”, “Progressing”

StringCondition applies to string-type attributes.

NumberCondition applies to numeric attributes (int, long, float, double).

BooleanCondition applies to boolean attributes.

JsonCondition applies to attributes that are JSON-encoded strings.

Operation can be:

• “exists”: Checks that the boolean attribute is present.
• ”equals”: Checks if the attribute exactly matches ExpectedValue.

ExpectedValue is required only for the “equals” operation.

ExpectedValue is only used if Operation == “equals”.

Represents the observations of a DeleteAttribute’s current state. Known .status.conditions.type are: “Available”, “Progressing”

FallbackSamplingRatio determines the percentage (0–100) of non-error traces that should be sampled. Error traces are always sampled.

ActionName is an optional identifier for this sampler rule. It can be used for referencing, labeling, or displaying the rule in UIs.

Notes provides free-form documentation or context for the user.

Disabled indicates whether the sampler is currently active. When true, the sampler will not be evaluated or applied.

Signals specifies the types of telemetry data this sampler should apply to. Typically, this includes “traces”, but may also include “logs” or “metrics”.

FallbackSamplingRatio determines the percentage (0–100) of non-error traces that should be sampled. Error traces are always sampled.

Conditions captures the current operational state of the sampler.

HttpRoute is the route name (from span attribute “http.route”) that this rule applies to.

ServiceName specifies the service that must emit the span for this rule to apply. Matches the value of the “service.name” attribute in the span.

MinimumLatencyThreshold is the latency in milliseconds that spans must exceed to be considered for sampling. Spans with latency >= this value are eligible.

FallbackSamplingRatio is the percentage (0–100) of traces to sample if the route and service match but the span latency is below the threshold.

FractionToRecord determines the percentage (0–100) of health checks traces to record. should be in range [0, 1] 0 (default) means no health checks traces will be recorded 1 means all health checks traces will be recorded @deprecated: use odigos config to enable/disable this behavior instead.

Operation defines the evaluation logic applied to the JSON value.

Supported values:

• is_valid_json
• is_invalid_json
• equals
• not_equals
• jsonpath_exists
• key_equals
• key_not_equals

JsonPath is required for:

• jsonpath_exists
• key_equals
• key_not_equals

It is a JSONPath expression used to navigate the JSON structure. Example: $.user.role or$.events[0].code

ExpectedValue is required for:

• equals
• not_equals
• key_equals
• key_not_equals

The value to compare against during evaluation.

The annotation name to be extracted. e.g. “kubectl.kubernetes.io/restartedAt”

The attribute key to be used for the resource attribute created from the annotation. e.g. “kubectl.kubernetes.restartedAt”

The source of the annotation. e.g. “pod” or “namespace” Deprecated: Use FromSources instead for specifying multiple sources with precedence.

The sources from which to extract the annotation, in order of precedence (most specific first). When multiple sources are specified, the most specific source (e.g., pod) takes precedence over less specific sources (e.g., namespace). If an annotation exists in multiple sources, the value from the most specific source will be used. Supported sources: “pod”, “namespace”, “node”

Collect the following container related attributes: k8s.container.name container.id container.image.name container.image.tag

collect replicaset related attributes (when relevant, e.g. for deployments): k8s.replicaset.name if CollectWorkloadUID is set, also collect: k8s.replicaset.uid DEPRECATED: ReplicaSet attributes are now collected by default during instrumentation.

Collect the following workload UID attributes: k8s.deployment.uid k8s.daemonset.uid k8s.statefulset.uid DEPRECATED: Workload UID attributes are now collected by default during instrumentation.

Collect the k8s.cluster.uid attribute, which is set to the uid of the namespace “kube-system”

list of labels to be extracted from the pod, and the attribute key to be used for the resource attribute created from each label.

list of annotations to be extracted from the pod, and the attribute key to be used for the resource attribute created from each annotation.

Collect the following container related attributes: k8s.container.name container.id container.image.name container.image.tag

collect replicaset related attributes (when relevant, e.g. for deployments): k8s.replicaset.name if CollectWorkloadUID is set, also collect: k8s.replicaset.uid DEPRECATED: ReplicaSet attributes are now collected by default during instrumentation.

Collect the following workload UID attributes: k8s.deployment.uid k8s.daemonset.uid k8s.statefulset.uid DEPRECATED: Workload UID attributes are now collected by default during instrumentation.

Collect the k8s.cluster.uid attribute, which is set to the uid of the namespace “kube-system”

list of labels to be extracted from the pod, and the attribute key to be used for the resource attribute created from each label.

list of annotations to be extracted from the pod, and the attribute key to be used for the resource attribute created from each annotation.

Represents the observations of a k8sattributes’ current state. Known .status.conditions.type are: “Available”, “Progressing”

The label name to be extracted. e.g. “app.kubernetes.io/name”

The attribute key to be used for the resource attribute created from the label. e.g. “app.kubernetes.name”

The source of the label. e.g. “pod” or “namespace” Deprecated: Use FromSources instead for specifying multiple sources with precedence.

The sources from which to extract the label, in order of precedence (most specific first). When multiple sources are specified, the most specific source (e.g., pod) takes precedence over less specific sources (e.g., namespace). If a label exists in multiple sources, the value from the most specific source will be used. Supported sources: “pod”, “namespace”, “node”

EndpointsFilters defines the list of route-based latency sampling filters. Each filter targets a specific service and HTTP route with a latency threshold.

ActionName is a user-defined identifier for this sampling action. It can be used to reference this policy in UIs or configuration tools.

Notes is an optional field for storing human-readable documentation or context for this sampler.

Disabled indicates whether the sampler is currently active. When true, this sampler will not be applied.

Signals lists the observability signal types (e.g., traces, metrics, logs) that this sampler applies to.

EndpointsFilters defines the list of route-based latency sampling filters. Each filter targets a specific service and HTTP route with a latency threshold.

Conditions contains the current status conditions for this sampler. Typical types include “Available” and “Progressing”.

Operation determines the numeric comparison to perform.

Valid operations:

• “exists”: Checks that the numeric attribute is present (non-null).
• “equals”: Checks if the attribute equals ExpectedValue.
• ”not_equals”: Checks if the attribute does not equal ExpectedValue.
• ”greater_than”: Checks if attribute > ExpectedValue.
• ”less_than”: Checks if attribute < ExpectedValue.
• ”greater_than_or_equal”: Checks if attribute >= ExpectedValue.
• ”less_than_or_equal”: Checks if attribute <= ExpectedValue.

For operations other than “exists”, ExpectedValue must be specified.

ExpectedValue is required for all operations except “exists”.

the name of the attribute to insert

if the value is a string, this field should be used. empty string is a valid value

Represents the observations of a piiMasking’s current state. Known .status.conditions.type are: “Available”, “Progressing”

SamplingPercentage determines the percentage (0–100) of traces to sample.

Represents the observations of a ProbabilisticSampler’s current state. Known .status.conditions.type are: “Available”, “Progressing”

Represents the observations of a RenameAttribute’s current state. Known .status.conditions.type are: “Available”, “Progressing”

ErrorSamplerSpec defines the desired state of ErrorSampler

LatencySamplerSpec defines the desired state of LatencySampler

ServiceNameSamplerSpec defines the desired state of ServiceNameSampler

SpanAttributeSamplerSpec defines the desired state of SpanAttributeSampler

ProbabilisticSamplerSpec defines the desired state of ProbabilisticSampler

automatically sample out health checks by a set fraction. health checks are discovered based on the livenessProbe and readinessProbe in the workload spec. if set, health checks are sampled regardless of any other configuration (head sampling at agent level, before traces are pushed into the collector pipeline) @deprecated: use odigos config to enable/disable this behavior instead.

ServiceName specifies the name of the service to look for within a trace. If any span in the trace comes from this service, the rule will apply.

SamplingRatio determines the percentage (0–100) of traces to sample when the specified service is present in the trace.

For example, a value of 100 means all such traces will be kept, while a value of 0 means all will be dropped.

FallbackSamplingRatio is the percentage (0–100) of traces to sample if the specified service is not present in the trace.

ServicesNameFilters defines rules for sampling traces based on the presence of specific service names. If a trace contains a span from one of the listed services, the associated sampling ratio is applied.

ActionName is an optional label to identify this sampling policy. It can be used for display purposes or integration with other tools.

Notes allows attaching additional free-form documentation or context to this sampler.

Disabled indicates whether this sampler should be active. If true, the sampler will not be applied.

Signals specifies which types of telemetry data this sampler applies to. Common values include “traces”, “metrics”, or “logs”.

ServicesNameFilters defines rules for sampling traces based on the presence of specific service names. If a trace contains a span from one of the listed services, the associated sampling ratio is applied.

Conditions is a list of status conditions for this sampler, following the standard Kubernetes conventions.

ServiceName specifies which service this filter applies to. Only spans originating from the given service will be evaluated against this filter.

AttributeKey indicates which attribute on the span to evaluate.

Condition is the rule or expression that will be used to evaluate the attribute’s value. Exactly one of the condition types must be set:

• StringCondition
• NumberCondition
• BooleanCondition
• JsonCondition

SamplingRatio determines the percentage (0–100) of traces to sample when the specified attribute satisfies the filter.

For example, a value of 100 means all such traces will be kept, while a value of 0 means all will be dropped.

FallbackSamplingRatio is the percentage (0–100) of spans to sample when the condition does not explicitly match. For example, if set to 50, then half of non-matching spans would be sampled.

AttributeFilters defines a list of criteria to decide how spans should be sampled based on their attributes. At least one filter is required.

ActionName is the name of the sampling action. This may be used to describe the purpose or intent of this sampler, for documentation or reference within other tools or systems.

Notes provides additional, free-form information about this sampler, such as a reference to a ticket, a link, or usage guidelines.

Disabled, if set to true, indicates that this sampler should not be applied at runtime.

Signals indicates which ObservabilitySignal types this sampler applies to. For instance, this could include traces, metrics, logs, etc.

AttributeFilters defines a list of criteria to decide how spans should be sampled based on their attributes. At least one filter is required.

Conditions is a list of the latest available observations of this sampler’s state.

Operation determines what comparison or check should be performed.

The valid operations are:

• “exists”: Checks that the attribute is present (and not an empty string).
• “equals”: String equality comparison with ExpectedValue.
• ”not_equals”: String inequality comparison with ExpectedValue.
• ”contains”: Checks if the attribute contains ExpectedValue as a substring.
• ”not_contains”: Checks if the attribute does not contain ExpectedValue.
• ”regex”: Interprets ExpectedValue as a regular expression (RE2 syntax) and checks for a match within the attribute.

For operations other than “exists”, ExpectedValue must be provided.

ExpectedValue is required for all operations except “exists”. Its usage depends on the chosen Operation, e.g. it may represent an exact string to match, a substring, or a regular expression.