kubernetes

package
v1.3.7 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Jul 24, 2026 License: AGPL-3.0 Imports: 28 Imported by: 0

Documentation

Index

Constants

This section is empty.

Variables

This section is empty.

Functions

This section is empty.

Types

type IngestService added in v1.3.0 

type IngestService struct {
	// contains filtered or unexported fields
}

IngestService reconciles a Kubernetes topology snapshot reported by an agent (or by the server's own local runtime under the LocalAgent id) into the store. It implements agentserver.KubernetesTopologyHandler.

func NewIngestService added in v1.3.0 

func NewIngestService(store TopologyStore, logger *slog.Logger) *IngestService

NewIngestService wires the Kubernetes topology ingest service.

func (*IngestService) HandleAgentEvent added in v1.3.0 

func (s *IngestService) HandleAgentEvent(ctx context.Context, agentID string, ev *agentpb.KubernetesTopology) error

HandleAgentEvent reconciles a full Kubernetes topology snapshot reported by a remote agent over gRPC.

func (*IngestService) Reconcile added in v1.3.0 

func (s *IngestService) Reconcile(ctx context.Context, agentID string, snap TopologySnapshot) error

Reconcile hard-reconciles a domain snapshot for agentID into the store. Used directly by the server's local-runtime reconcile loop (under LocalAgent) and indirectly by HandleAgentEvent for remote agents.

func (*IngestService) SetBroadcaster added in v1.3.0 

func (s *IngestService) SetBroadcaster(fn func(eventType string, data any))

SetBroadcaster wires an SSE broadcaster. When set, a kubernetes.topology_changed event carrying the agent id is emitted after a remote agent's snapshot is reconciled, so connected clients scoped to that agent refetch.

type K8sAlertCallback added in v1.2.0 

type K8sAlertCallback func(evt alert.Event)

K8sAlertCallback is the function signature for emitting alert events.

type K8sAlertChecker added in v1.2.0 

type K8sAlertChecker struct {
	// contains filtered or unexported fields
}

K8sAlertChecker detects alert conditions in Kubernetes workloads and nodes.

func NewK8sAlertChecker added in v1.2.0 

func NewK8sAlertChecker(logger *slog.Logger) *K8sAlertChecker

NewK8sAlertChecker creates a new K8sAlertChecker.

func (*K8sAlertChecker) CheckCrashLoopBackOff added in v1.2.0 

func (c *K8sAlertChecker) CheckCrashLoopBackOff(pods []K8sPod)

CheckCrashLoopBackOff checks pods for crash-loop patterns based on restart counts. A pod is considered in crash-loop when it accumulates defaultCrashLoopRestarts or more restart increments within defaultCrashLoopWindow.

func (*K8sAlertChecker) CheckNodeConditions added in v1.2.0 

func (c *K8sAlertChecker) CheckNodeConditions(nodes []K8sNode)

CheckNodeConditions checks nodes for unhealthy conditions: NotReady, MemoryPressure, DiskPressure, PIDPressure.

func (*K8sAlertChecker) CheckWorkloadReplicas added in v1.2.0 

func (c *K8sAlertChecker) CheckWorkloadReplicas(workloads []K8sWorkload)

CheckWorkloadReplicas checks for workloads where ready < desired for longer than the grace period. Returns IDs of workloads that should alert.

func (*K8sAlertChecker) SetAlertCallback added in v1.2.0 

func (c *K8sAlertChecker) SetAlertCallback(cb K8sAlertCallback)

SetAlertCallback sets the callback for emitting alerts.

type K8sClusterOverview added in v1.2.0 

type K8sClusterOverview struct {
	NamespaceCount  int                   `json:"namespace_count"`
	NodeCount       int                   `json:"node_count"`
	NodeReadyCount  int                   `json:"node_ready_count"`
	PodStatus       K8sPodStatusBreakdown `json:"pod_status"`
	WorkloadCount   int                   `json:"workload_count"`
	WorkloadHealthy int                   `json:"workload_healthy"`
	ClusterHealth   string                `json:"cluster_health"` // healthy, degraded, unhealthy
	Namespaces      []K8sNamespaceSummary `json:"namespaces"`
}

K8sClusterOverview holds aggregated cluster state for the dashboard.

type K8sCondition added in v1.2.0 

type K8sCondition struct {
	Type           string
	Status         string // True, False, Unknown
	Reason         string
	Message        string
	LastTransition time.Time
}

K8sCondition represents a single Kubernetes condition.

type K8sContainerStatus added in v1.2.0 

type K8sContainerStatus struct {
	Name         string
	Image        string
	Ready        bool
	RestartCount int32
	State        string // running, waiting, terminated
	StateReason  string
	StartedAt    *time.Time
}

K8sContainerStatus holds per-container runtime state within a pod.

type K8sEvent added in v1.2.0 

type K8sEvent struct {
	Type      string
	Reason    string
	Message   string
	Source    string
	FirstSeen time.Time
	LastSeen  time.Time
	Count     int32
}

K8sEvent is a summarised Kubernetes event (v1.Event).

type K8sEventRef added in v1.3.0 

type K8sEventRef struct {
	K8sEvent
	InvolvedKind      string
	InvolvedNamespace string
	InvolvedName      string
}

K8sEventRef is a K8sEvent plus the object it concerns, so the server can store it per-agent and serve it back for the matching workload or pod detail view.

type K8sNamespaceSummary added in v1.2.0 

type K8sNamespaceSummary struct {
	Name          string `json:"name"`
	WorkloadCount int    `json:"workload_count"`
	PodCount      int    `json:"pod_count"`
	Healthy       bool   `json:"healthy"`
}

K8sNamespaceSummary holds per-namespace workload and pod counts.

type K8sNode added in v1.2.0 

type K8sNode struct {
	AgentID           string // reporting agent; empty/LocalAgent for the server's own cluster
	Name              string
	Roles             []string
	Conditions        []K8sCondition
	Status            string // ready, not-ready, unknown
	Capacity          K8sResourceQuantity
	Allocatable       K8sResourceQuantity
	RunningPods       int
	KubernetesVersion string
	OSImage           string
	Architecture      string
	CreatedAt         time.Time
}

K8sNode represents a Kubernetes cluster node.

type K8sPod added in v1.2.0 

type K8sPod struct {
	AgentID      string // reporting agent; empty/LocalAgent for the server's own cluster
	Name         string
	Namespace    string
	Status       string // Running, Pending, Succeeded, Failed, Unknown
	StatusReason string // CrashLoopBackOff, ImagePullBackOff, etc.
	RestartCount int32
	NodeName     string
	PodIP        string
	HostIP       string
	Containers   []K8sContainerStatus
	WorkloadRef  string // owning workload ID
	CreatedAt    time.Time
}

K8sPod represents a single Kubernetes pod.

type K8sPodStatusBreakdown added in v1.2.0 

type K8sPodStatusBreakdown struct {
	Running   int `json:"running"`
	Pending   int `json:"pending"`
	Failed    int `json:"failed"`
	Succeeded int `json:"succeeded"`
	Unknown   int `json:"unknown"`
}

K8sPodStatusBreakdown counts pods by phase.

type K8sResourceQuantity added in v1.2.0 

type K8sResourceQuantity struct {
	CPUMillicores int64
	MemoryBytes   int64
	Pods          int64
}

K8sResourceQuantity holds parsed resource capacity values.

type K8sWorkload added in v1.2.0 

type K8sWorkload struct {
	ID              string // "{namespace}/{kind}/{name}"
	AgentID         string // reporting agent; empty/LocalAgent for the server's own cluster
	Name            string
	Namespace       string
	Kind            string // Deployment, StatefulSet, DaemonSet, Job
	Images          []string
	ReadyReplicas   int32
	DesiredReplicas int32
	Status          string // healthy, degraded, progressing, failed
	Conditions      []K8sCondition
	Labels          map[string]string
	CreatedAt       time.Time
	LastTransition  time.Time
}

K8sWorkload represents a Kubernetes controller.

type K8sWorkloadGroup added in v1.2.0 

type K8sWorkloadGroup struct {
	Namespace string
	Workloads []K8sWorkload
}

K8sWorkloadGroup groups workloads by namespace.

type NamespaceFilter 

type NamespaceFilter struct {
	// contains filtered or unexported fields
}

NamespaceFilter controls which namespaces maintenant monitors.

func NewNamespaceFilter 

func NewNamespaceFilter(allowCSV, excludeCSV string) *NamespaceFilter

NewNamespaceFilter creates a filter from env var values. allowCSV = MAINTENANT_K8S_NAMESPACES (comma-separated allowlist). excludeCSV = MAINTENANT_K8S_EXCLUDE_NAMESPACES (comma-separated blocklist, appended to defaults).

func (*NamespaceFilter) IsAllowed 

func (f *NamespaceFilter) IsAllowed(namespace string) bool

IsAllowed returns true if the namespace should be monitored.

type NodeResourceMetrics added in v1.2.0 

type NodeResourceMetrics struct {
	Name                  string
	CPUMillicores         int64
	CPUCapacityMillicores int64
	MemBytes              int64
	MemCapacityBytes      int64
	Timestamp             time.Time
}

NodeResourceMetrics holds resource metrics for a single node.

type PodFilters added in v1.2.0 

type PodFilters struct {
	Workload string // owning workload ID prefix
	Node     string // node name
	Status   string // Running, Pending, etc.
}

PodFilters are optional filters for ListPods.

type PodResourceMetrics added in v1.2.0 

type PodResourceMetrics struct {
	Name          string
	Namespace     string
	CPUMillicores int64
	MemBytes      int64
	MemLimitBytes int64
	Timestamp     time.Time
}

PodResourceMetrics holds aggregated resource metrics for a single pod.

type Runtime 

type Runtime struct {
	// contains filtered or unexported fields
}

Runtime implements runtime.Runtime for Kubernetes.

func NewRuntime 

func NewRuntime(logger *slog.Logger, nsFilter *NamespaceFilter) (*Runtime, error)

NewRuntime creates a Kubernetes runtime. Connection is deferred to Connect().

func (*Runtime) Close 

func (r *Runtime) Close() error

func (*Runtime) ClusterOverview added in v1.2.0 

func (r *Runtime) ClusterOverview(ctx context.Context) (*K8sClusterOverview, error)

ClusterOverview aggregates cluster-wide state from namespaces, nodes, workloads, and pods into a single overview structure.

func (*Runtime) Connect 

func (r *Runtime) Connect(ctx context.Context) error

func (*Runtime) DiscoverAll 

func (r *Runtime) DiscoverAll(ctx context.Context) ([]*cmodel.Container, error)

func (*Runtime) FetchLogSnippet 

func (r *Runtime) FetchLogSnippet(ctx context.Context, externalID string) (string, error)

FetchLogSnippet retrieves the last 50 lines for die event snippets. Satisfies container.LogFetcher interface.

func (*Runtime) FetchLogs 

func (r *Runtime) FetchLogs(ctx context.Context, externalID string, lines int, timestamps bool) ([]string, error)

func (*Runtime) GetHealthInfo 

func (r *Runtime) GetHealthInfo(ctx context.Context, externalID string) (*runtime.HealthInfo, error)

func (*Runtime) GetNodeMetrics added in v1.2.0 

func (r *Runtime) GetNodeMetrics(ctx context.Context, name string) (*NodeResourceMetrics, error)

GetNodeMetrics queries metrics-server for a node's CPU and memory usage.

func (*Runtime) GetPodDetail added in v1.2.0 

func (r *Runtime) GetPodDetail(ctx context.Context, namespace, name string) (*K8sPod, []K8sEvent, error)

GetPodDetail returns details for a single pod plus recent events.

func (*Runtime) GetPodMetrics added in v1.2.0 

func (r *Runtime) GetPodMetrics(ctx context.Context, namespace, name string) (*PodResourceMetrics, error)

GetPodMetrics queries metrics-server for a pod's CPU and memory usage.

func (*Runtime) GetWorkload added in v1.2.0 

func (r *Runtime) GetWorkload(ctx context.Context, id string) (*K8sWorkload, []K8sPod, []K8sEvent, error)

GetWorkload returns the detail for a single workload, its pods, and recent events. id format: "namespace/Kind/name".

func (*Runtime) IsConnected 

func (r *Runtime) IsConnected() bool

func (*Runtime) ListAllEvents added in v1.3.0 

func (r *Runtime) ListAllEvents(ctx context.Context) ([]K8sEventRef, error)

ListAllEvents returns recent events across all allowed namespaces, each tagged with the object it concerns, so the server can persist them per-agent and serve them back on workload/pod detail views.

func (*Runtime) ListContainerNames 

func (r *Runtime) ListContainerNames(ctx context.Context, externalID string) ([]string, error)

ListContainerNames returns the container names in a workload's pod spec. For controllers, resolves to a pod's spec. For bare pods, reads the pod directly.

func (*Runtime) ListNamespaces added in v1.2.0 

func (r *Runtime) ListNamespaces(ctx context.Context) ([]string, error)

ListNamespaces returns the allowed namespace names from the cluster. The result respects the allowlist/blocklist configured via env vars.

func (*Runtime) ListNodes added in v1.2.0 

func (r *Runtime) ListNodes(ctx context.Context) ([]K8sNode, error)

ListNodes returns all cluster nodes with their resource capacity and status.

func (*Runtime) ListPods added in v1.2.0 

func (r *Runtime) ListPods(ctx context.Context, namespaces []string, filters PodFilters) ([]K8sPod, error)

ListPods returns a flat pod list optionally filtered by workload, node, and status.

func (*Runtime) ListWorkloads added in v1.2.0 

func (r *Runtime) ListWorkloads(ctx context.Context, namespaces []string) ([]K8sWorkloadGroup, error)

ListWorkloads returns workloads grouped by namespace. When namespaces is non-empty only those namespaces are queried; otherwise all allowed namespaces are included.

func (*Runtime) MetricsAvailable added in v1.2.0 

func (r *Runtime) MetricsAvailable() bool

MetricsAvailable reports whether the metrics-server API was detected on startup.

func (*Runtime) Name 

func (r *Runtime) Name() string

func (*Runtime) SetDisconnected 

func (r *Runtime) SetDisconnected()

func (*Runtime) StatsSnapshot 

func (r *Runtime) StatsSnapshot(ctx context.Context, externalID string) (*runtime.RawStats, error)

func (*Runtime) StreamEvents 

func (r *Runtime) StreamEvents(ctx context.Context) <-chan runtime.RuntimeEvent

func (*Runtime) StreamLogs 

func (r *Runtime) StreamLogs(ctx context.Context, externalID string, lines int, timestamps bool) (io.ReadCloser, error)

func (*Runtime) TryConnect added in v1.2.14 

func (r *Runtime) TryConnect(ctx context.Context) error

type SnapshotSource added in v1.3.0 

type SnapshotSource interface {
	ListNamespaces(ctx context.Context) ([]string, error)
	ListWorkloads(ctx context.Context, namespaces []string) ([]K8sWorkloadGroup, error)
	ListPods(ctx context.Context, namespaces []string, filters PodFilters) ([]K8sPod, error)
	ListNodes(ctx context.Context) ([]K8sNode, error)
	ListAllEvents(ctx context.Context) ([]K8sEventRef, error)
}

SnapshotSource is the subset of the Kubernetes runtime needed to build a topology snapshot. *Runtime satisfies it.

type TopologySnapshot added in v1.3.0 

type TopologySnapshot struct {
	Namespaces []string
	Workloads  []K8sWorkload
	Pods       []K8sPod
	Nodes      []K8sNode
	Events     []K8sEventRef
}

TopologySnapshot is a full point-in-time view of an agent's Kubernetes cluster. The server reconciles it against the rows already held for that agent, hard-deleting anything absent from the snapshot. Produced by the server's local runtime (under the LocalAgent id) and by remote agents (under their own id).

func SnapshotFromRuntime added in v1.3.0 

func SnapshotFromRuntime(ctx context.Context, src SnapshotSource) (TopologySnapshot, error)

SnapshotFromRuntime queries the live cluster for namespaces, workloads, pods and nodes. Nodes are best-effort: a cluster-scoped RBAC denial on nodes does not fail the whole snapshot (the agent may only be granted namespaced access).

type TopologyStore added in v1.3.0 

type TopologyStore interface {
	ReplaceNamespacesForAgent(ctx context.Context, agentID string, namespaces []string) error
	ReplaceWorkloadsForAgent(ctx context.Context, agentID string, workloads []K8sWorkload) error
	ReplacePodsForAgent(ctx context.Context, agentID string, pods []K8sPod) error
	ReplaceNodesForAgent(ctx context.Context, agentID string, nodes []K8sNode) error
	ReplaceEventsForAgent(ctx context.Context, agentID string, events []K8sEventRef) error
}

TopologyStore persists per-agent Kubernetes topology (SQLite-backed).

Source Files

View all Source files

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.