-Network Intelligence for Kubernetes
+Network Observability for SREs & AI Agents
Live Demo · Docs
@@ -17,9 +17,17 @@
---
-* **Cluster-wide, real-time visibility into every packet, API call, and service interaction.**
-* Replay any moment in time.
-* Resolve incidents at the speed of LLMs. 100% on-premises.
+Kubeshark captures cluster-wide network traffic at the speed and scale of Kubernetes, continuously, at the kernel level using eBPF. It consolidates a highly fragmented picture — dozens of nodes, thousands of workloads, millions of connections — into a single, queryable view with full Kubernetes and API context.
+
+Network data is available to **AI agents via [MCP](https://docs.kubeshark.com/en/mcp)** and to **human operators via a [dashboard](https://docs.kubeshark.com/en/v2)**.
+
+**What's captured, cluster-wide:**
+
+- **L4 Packets & TCP Metrics** — retransmissions, RTT, window saturation, connection lifecycle, packet loss across every node-to-node path ([TCP insights →](https://docs.kubeshark.com/en/mcp/tcp_insights))
+- **L7 API Calls** — real-time request/response matching with full payload parsing: HTTP, gRPC, GraphQL, Redis, Kafka, DNS ([API dissection →](https://docs.kubeshark.com/en/v2/l7_api_dissection))
+- **Decrypted TLS** — eBPF-based TLS decryption without key management
+- **Kubernetes Context** — every packet and API call resolved to pod, service, namespace, and node
+- **PCAP Retention** — point-in-time raw packet snapshots, exportable for Wireshark ([Snapshots →](https://docs.kubeshark.com/en/v2/traffic_snapshots))
@@ -34,33 +42,37 @@ helm install kubeshark kubeshark/kubeshark
Dashboard opens automatically. You're capturing traffic.
-**With AI** — connect your assistant and debug with natural language:
+**Connect an AI agent** via MCP:
```bash
brew install kubeshark
claude mcp add kubeshark -- kubeshark mcp
```
+[MCP setup guide →](https://docs.kubeshark.com/en/mcp)
+
+---
+
+### AI-Powered Network Analysis
+
+Kubeshark exposes all cluster-wide network data via MCP (Model Context Protocol). AI agents can query L4 metrics, investigate L7 API calls, analyze traffic patterns, and run root cause analysis — through natural language. Use cases include incident response, root cause analysis, troubleshooting, debugging, and reliability workflows.
+
> *"Why did checkout fail at 2:15 PM?"*
> *"Which services have error rates above 1%?"*
+> *"Show TCP retransmission rates across all node-to-node paths"*
+> *"Trace request abc123 through all services"*
+
+Works with Claude Code, Cursor, and any MCP-compatible AI.
+
+
[MCP setup guide →](https://docs.kubeshark.com/en/mcp)
---
-## Why Kubeshark
+### L7 API Dissection
-- **Instant root cause** — trace requests across services, see exact errors
-- **Zero instrumentation** — no code changes, no SDKs, just deploy
-- **Full payload capture** — request/response bodies, headers, timing
-- **TLS decryption** — see encrypted traffic without managing keys
-- **AI-ready** — query traffic with natural language via MCP
-
----
-
-### Traffic Analysis and API Dissection
-
-Capture and inspect every API call across your cluster—HTTP, gRPC, Redis, Kafka, DNS, and more. Request/response matching with full payloads, parsed according to protocol specifications. Headers, timing, and complete context. Zero instrumentation required.
+Cluster-wide request/response matching with full payloads, parsed according to protocol specifications. HTTP, gRPC, Redis, Kafka, DNS, and more. Every API call resolved to source and destination pod, service, namespace, and node. No code instrumentation required.
@@ -68,27 +80,15 @@ Capture and inspect every API call across your cluster—HTTP, gRPC, Redis, Kafk
### L4/L7 Workload Map
-Visualize how your services communicate. See dependencies, traffic flow, and identify anomalies at a glance.
+Cluster-wide view of service communication: dependencies, traffic flow, and anomalies across all nodes and namespaces.
[Learn more →](https://docs.kubeshark.com/en/v2/service_map)
-### AI-Powered Root Cause Analysis
-
-Resolve production issues in minutes instead of hours. Connect your AI assistant and investigate incidents using natural language. Build network-aware AI agents for forensics, monitoring, compliance, and security.
-
-> *"Why did checkout fail at 2:15 PM?"*
-> *"Which services have error rates above 1%?"*
-> *"Trace request abc123 through all services"*
-
-Works with Claude Code, Cursor, and any MCP-compatible AI.
-
-[MCP setup guide →](https://docs.kubeshark.com/en/mcp)
-
### Traffic Retention
-Retain every packet. Take snapshots. Export PCAP files. Replay any moment in time.
+Continuous raw packet capture with point-in-time snapshots. Export PCAP files for offline analysis with Wireshark or other tools.
@@ -105,7 +105,7 @@ Retain every packet. Take snapshots. Export PCAP files. Replay any moment in tim
| [**L7 API Dissection**](https://docs.kubeshark.com/en/v2/l7_api_dissection) | Request/response matching with full payloads and protocol parsing |
| [**Protocol Support**](https://docs.kubeshark.com/en/protocols) | HTTP, gRPC, GraphQL, Redis, Kafka, DNS, and more |
| [**TLS Decryption**](https://docs.kubeshark.com/en/encrypted_traffic) | eBPF-based decryption without key management |
-| [**AI-Powered Analysis**](https://docs.kubeshark.com/en/v2/ai_powered_analysis) | Query traffic with Claude, Cursor, or any MCP-compatible AI |
+| [**AI-Powered Analysis**](https://docs.kubeshark.com/en/v2/ai_powered_analysis) | Query cluster-wide network data with Claude, Cursor, or any MCP-compatible AI |
| [**Display Filters**](https://docs.kubeshark.com/en/v2/kfl2) | Wireshark-inspired display filters for precise traffic analysis |
| [**100% On-Premises**](https://docs.kubeshark.com/en/air_gapped) | Air-gapped support, no external dependencies |
diff --git a/config/configStructs/tapConfig.go b/config/configStructs/tapConfig.go
index 3dbb30b1a..e0b6ddfbb 100644
--- a/config/configStructs/tapConfig.go
+++ b/config/configStructs/tapConfig.go
@@ -330,6 +330,12 @@ type SnapshotsCloudAzblobConfig struct {
StorageKey string `yaml:"storageKey" json:"storageKey" default:""`
}
+type SnapshotsCloudGCSConfig struct {
+ Bucket string `yaml:"bucket" json:"bucket" default:""`
+ Project string `yaml:"project" json:"project" default:""`
+ CredentialsJson string `yaml:"credentialsJson" json:"credentialsJson" default:""`
+}
+
type SnapshotsCloudConfig struct {
Provider string `yaml:"provider" json:"provider" default:""`
Prefix string `yaml:"prefix" json:"prefix" default:""`
@@ -337,6 +343,7 @@ type SnapshotsCloudConfig struct {
Secrets []string `yaml:"secrets" json:"secrets" default:"[]"`
S3 SnapshotsCloudS3Config `yaml:"s3" json:"s3"`
Azblob SnapshotsCloudAzblobConfig `yaml:"azblob" json:"azblob"`
+ GCS SnapshotsCloudGCSConfig `yaml:"gcs" json:"gcs"`
}
type SnapshotsConfig struct {
diff --git a/helm-chart/README.md b/helm-chart/README.md
index 6a90917be..12c186153 100644
--- a/helm-chart/README.md
+++ b/helm-chart/README.md
@@ -146,10 +146,10 @@ Example for overriding image names:
| `tap.capture.dbMaxSize` | Maximum size for capture database (e.g., `4Gi`, `2000Mi`). | `500Mi` |
| `tap.snapshots.local.storageClass` | Storage class for local snapshots volume. When empty, uses `emptyDir`. When set, creates a PVC with this storage class | `""` |
| `tap.snapshots.local.storageSize` | Storage size for local snapshots volume (supports K8s quantity format: `1Gi`, `500Mi`, etc.) | `20Gi` |
-| `tap.snapshots.cloud.provider` | Cloud storage provider for snapshots: `s3` or `azblob`. Empty string disables cloud storage. See [Cloud Storage docs](docs/snapshots_cloud_storage.md). | `""` |
+| `tap.snapshots.cloud.provider` | Cloud storage provider for snapshots: `s3`, `azblob`, or `gcs`. Empty string disables cloud storage. See [Cloud Storage docs](docs/snapshots_cloud_storage.md). | `""` |
| `tap.snapshots.cloud.prefix` | Key prefix in the bucket/container (e.g. `snapshots/`). See [Cloud Storage docs](docs/snapshots_cloud_storage.md). | `""` |
-| `tap.snapshots.cloud.configMaps` | Names of pre-existing ConfigMaps with cloud storage env vars. Alternative to inline `s3`/`azblob` values below. See [Cloud Storage docs](docs/snapshots_cloud_storage.md). | `[]` |
-| `tap.snapshots.cloud.secrets` | Names of pre-existing Secrets with cloud storage credentials. Alternative to inline `s3`/`azblob` values below. See [Cloud Storage docs](docs/snapshots_cloud_storage.md). | `[]` |
+| `tap.snapshots.cloud.configMaps` | Names of pre-existing ConfigMaps with cloud storage env vars. Alternative to inline `s3`/`azblob`/`gcs` values below. See [Cloud Storage docs](docs/snapshots_cloud_storage.md). | `[]` |
+| `tap.snapshots.cloud.secrets` | Names of pre-existing Secrets with cloud storage credentials. Alternative to inline `s3`/`azblob`/`gcs` values below. See [Cloud Storage docs](docs/snapshots_cloud_storage.md). | `[]` |
| `tap.snapshots.cloud.s3.bucket` | S3 bucket name. When set, the chart auto-creates a ConfigMap with `SNAPSHOT_AWS_BUCKET`. | `""` |
| `tap.snapshots.cloud.s3.region` | AWS region for the S3 bucket. | `""` |
| `tap.snapshots.cloud.s3.accessKey` | AWS access key ID. When set, the chart auto-creates a Secret with `SNAPSHOT_AWS_ACCESS_KEY`. | `""` |
@@ -159,6 +159,9 @@ Example for overriding image names:
| `tap.snapshots.cloud.azblob.storageAccount` | Azure storage account name. When set, the chart auto-creates a ConfigMap with `SNAPSHOT_AZBLOB_STORAGE_ACCOUNT`. | `""` |
| `tap.snapshots.cloud.azblob.container` | Azure blob container name. | `""` |
| `tap.snapshots.cloud.azblob.storageKey` | Azure storage account access key. When set, the chart auto-creates a Secret with `SNAPSHOT_AZBLOB_STORAGE_KEY`. | `""` |
+| `tap.snapshots.cloud.gcs.bucket` | GCS bucket name. When set, the chart auto-creates a ConfigMap with `SNAPSHOT_GCS_BUCKET`. | `""` |
+| `tap.snapshots.cloud.gcs.project` | GCP project ID. | `""` |
+| `tap.snapshots.cloud.gcs.credentialsJson` | Service account JSON key. When set, the chart auto-creates a Secret with `SNAPSHOT_GCS_CREDENTIALS_JSON`. | `""` |
| `tap.delayedDissection.cpu` | CPU allocation for delayed dissection jobs | `1` |
| `tap.delayedDissection.memory` | Memory allocation for delayed dissection jobs | `4Gi` |
| `tap.release.repo` | URL of the Helm chart repository | `https://helm.kubeshark.com` |
diff --git a/helm-chart/docs/snapshots_cloud_storage.md b/helm-chart/docs/snapshots_cloud_storage.md
index 50997a68e..09c6b9745 100644
--- a/helm-chart/docs/snapshots_cloud_storage.md
+++ b/helm-chart/docs/snapshots_cloud_storage.md
@@ -2,7 +2,7 @@
Kubeshark can upload and download snapshots to cloud object storage, enabling cross-cluster sharing, backup/restore, and long-term retention.
-Supported providers: **Amazon S3** (`s3`) and **Azure Blob Storage** (`azblob`).
+Supported providers: **Amazon S3** (`s3`), **Azure Blob Storage** (`azblob`), and **Google Cloud Storage** (`gcs`).
## Helm Values
@@ -10,7 +10,7 @@ Supported providers: **Amazon S3** (`s3`) and **Azure Blob Storage** (`azblob`).
tap:
snapshots:
cloud:
- provider: "" # "s3" or "azblob" (empty = disabled)
+ provider: "" # "s3", "azblob", or "gcs" (empty = disabled)
prefix: "" # key prefix in the bucket/container (e.g. "snapshots/")
configMaps: [] # names of pre-existing ConfigMaps with cloud config env vars
secrets: [] # names of pre-existing Secrets with cloud credentials
@@ -25,6 +25,10 @@ tap:
storageAccount: ""
container: ""
storageKey: ""
+ gcs:
+ bucket: ""
+ project: ""
+ credentialsJson: ""
```
- `provider` selects which cloud backend to use. Leave empty to disable cloud storage.
@@ -278,3 +282,212 @@ tap:
secrets:
- kubeshark-azblob-creds
```
+
+---
+
+## Google Cloud Storage
+
+### Environment Variables
+
+| Variable | Required | Description |
+|----------|----------|-------------|
+| `SNAPSHOT_GCS_BUCKET` | Yes | GCS bucket name |
+| `SNAPSHOT_GCS_PROJECT` | No | GCP project ID |
+| `SNAPSHOT_GCS_CREDENTIALS_JSON` | No | Service account JSON key (empty = use Application Default Credentials) |
+| `SNAPSHOT_CLOUD_PREFIX` | No | Key prefix in the bucket (e.g. `snapshots/`) |
+
+### Authentication Methods
+
+Credentials are resolved in this order:
+
+1. **Service Account JSON Key** -- If `SNAPSHOT_GCS_CREDENTIALS_JSON` is set, the provided JSON key is used directly.
+2. **Application Default Credentials** -- When no JSON key is provided, the GCP SDK default credential chain is used:
+ - **Workload Identity** (GKE pod identity) -- recommended for production on GKE
+ - GCE instance metadata (Compute Engine default service account)
+ - Standard GCP environment variables (`GOOGLE_APPLICATION_CREDENTIALS`)
+ - `gcloud` CLI credentials
+
+The provider validates bucket access on startup via `Bucket.Attrs()`. If the bucket is inaccessible, the hub will fail to start.
+
+### Required IAM Permissions
+
+The service account needs different IAM roles depending on the access level:
+
+**Read-only** (download, list, and sync snapshots from cloud):
+
+| Role | Permissions provided | Purpose |
+|------|---------------------|---------|
+| `roles/storage.legacyBucketReader` | `storage.buckets.get`, `storage.objects.list` | Hub startup (bucket validation) + listing snapshots |
+| `roles/storage.objectViewer` | `storage.objects.get`, `storage.objects.list` | Downloading snapshots, checking existence, reading metadata |
+
+```bash
+gcloud storage buckets add-iam-policy-binding gs://BUCKET_NAME \
+ --member="serviceAccount:SA_EMAIL" \
+ --role="roles/storage.legacyBucketReader"
+gcloud storage buckets add-iam-policy-binding gs://BUCKET_NAME \
+ --member="serviceAccount:SA_EMAIL" \
+ --role="roles/storage.objectViewer"
+```
+
+**Read-write** (upload and delete snapshots in addition to read):
+
+Add `roles/storage.objectAdmin` instead of `roles/storage.objectViewer` to also grant `storage.objects.create` and `storage.objects.delete`:
+
+| Role | Permissions provided | Purpose |
+|------|---------------------|---------|
+| `roles/storage.legacyBucketReader` | `storage.buckets.get`, `storage.objects.list` | Hub startup (bucket validation) + listing snapshots |
+| `roles/storage.objectAdmin` | `storage.objects.*` | Full object CRUD (upload, download, delete, list, metadata) |
+
+```bash
+gcloud storage buckets add-iam-policy-binding gs://BUCKET_NAME \
+ --member="serviceAccount:SA_EMAIL" \
+ --role="roles/storage.legacyBucketReader"
+gcloud storage buckets add-iam-policy-binding gs://BUCKET_NAME \
+ --member="serviceAccount:SA_EMAIL" \
+ --role="roles/storage.objectAdmin"
+```
+
+### Example: Inline Values (simplest approach)
+
+```yaml
+tap:
+ snapshots:
+ cloud:
+ provider: "gcs"
+ gcs:
+ bucket: my-kubeshark-snapshots
+ project: my-gcp-project
+```
+
+Or with a service account key via `--set`:
+
+```bash
+helm install kubeshark kubeshark/kubeshark \
+ --set tap.snapshots.cloud.provider=gcs \
+ --set tap.snapshots.cloud.gcs.bucket=my-kubeshark-snapshots \
+ --set tap.snapshots.cloud.gcs.project=my-gcp-project \
+ --set-file tap.snapshots.cloud.gcs.credentialsJson=service-account.json
+```
+
+### Example: Workload Identity (recommended for GKE)
+
+Create a ConfigMap with bucket configuration:
+
+```yaml
+apiVersion: v1
+kind: ConfigMap
+metadata:
+ name: kubeshark-gcs-config
+data:
+ SNAPSHOT_GCS_BUCKET: my-kubeshark-snapshots
+ SNAPSHOT_GCS_PROJECT: my-gcp-project
+```
+
+Set Helm values:
+
+```yaml
+tap:
+ snapshots:
+ cloud:
+ provider: "gcs"
+ configMaps:
+ - kubeshark-gcs-config
+```
+
+Configure GKE Workload Identity to allow the Kubernetes service account to impersonate the GCP service account:
+
+```bash
+# Ensure the GKE cluster has Workload Identity enabled
+# (--workload-pool=PROJECT_ID.svc.id.goog at cluster creation)
+
+# Create a GCP service account (if not already created)
+gcloud iam service-accounts create kubeshark-gcs \
+ --display-name="Kubeshark GCS Snapshots"
+
+# Grant bucket access (read-write — see Required IAM Permissions above)
+gcloud storage buckets add-iam-policy-binding gs://BUCKET_NAME \
+ --member="serviceAccount:kubeshark-gcs@PROJECT_ID.iam.gserviceaccount.com" \
+ --role="roles/storage.legacyBucketReader"
+gcloud storage buckets add-iam-policy-binding gs://BUCKET_NAME \
+ --member="serviceAccount:kubeshark-gcs@PROJECT_ID.iam.gserviceaccount.com" \
+ --role="roles/storage.objectAdmin"
+
+# Allow the K8s service account to impersonate the GCP service account
+# Note: the K8s SA name is "