Tentacles

CI/CD and GitOps that reach into every cluster.

One control plane for builds and deploys — agents execute where your workloads live, from VAS to customer clusters, without juggling Jenkins and Argo side by side.

Pipelines — build, GitOps apply, and rollout strategy in one execution timeline.

Product

Replace the Jenkins + Argo handoff with one pipeline.

Tentacles owns the path from source to healthy pods: secrets, OCI build, GitOps apply, and rollout strategy — visible in one execution timeline.

  • OCI build step

    Build and push container images from the agent — no Jenkins job YAML to babysit.

  • GitOps deploy

    Clone manifests, server-side apply, wait for healthy, optional prune and rollback.

  • Rollout strategies

    Rolling, canary weight steps, or blue/green traffic switch from one pipeline spec.

  • Multi-cluster agents

    Reach into customer, VAS, and edge clusters through agents that dial home over gRPC.

  • Officeless promotions

    Promote Visual Programming, Data Manager, and Workflows from Dev to PRD with gateway sync and a -uat2 rewrite on deploy.

  • Policy gates

    Built-in and Rego checks block risky manifests — no :latest, no privileged containers — before GitOps ever applies them.

Feature guides

Every surface, with a step-by-step walkthrough.

From pipelines and GitOps to Officeless Visual Programming, Data Manager, and Workflows — click any card for previews and a guided walkthrough.

Open guide →
Core

Live pipeline dashboard

Total pipelines, running executions, success rate, and online agents in one view, backed by a 7-day execution chart and recent run history.

Open guide →
Core

Pipelines

Compose an OCI build step, a GitOps apply, and a rollout strategy into one DeploymentPipeline spec — rolling, canary, or blue/green.

Open guide →
Core

Executions

Full run history across every pipeline with status, start time, image digest, and error detail — drill into any run's live stage timeline.

Open guide →
Core

GitOps apply

Clone manifests, server-side apply, and wait for healthy Deployments, StatefulSets, DaemonSets, and Pods before the step succeeds.

Open guide →
Core

Agents

Lightweight agents dial home over gRPC to build images and apply manifests inside customer, VAS, or edge clusters — least-privilege by design.

Open guide →
Officeless

Officeless Visual Programming

Sync VP apps from the Dev gateway, select which app ids to promote, and deploy to PRD with an automatic -uat2 host rewrite.

Open guide →
Officeless

Officeless Data Manager

Promote Data Manager table and workflow definitions from Dev to PRD with gateway config resolved per environment on deploy.

Open guide →
Officeless

Officeless Workflows

Move Workflow automations from Dev into PRD with dependency-aware sync and an environment rewrite — no manual export or import.

Open guide →
Ops

Notifications

Route pipeline and deploy events to Slack, email, Google Chat, or webhooks so the right channel sees success, failure, and canary pauses.

Open guide →
Security

Secret providers

Register Vault, AWS Secrets Manager, GCP Secret Manager, or Kubernetes secret providers once; agents resolve values only where a pipeline runs.

Open guide →
Ops

Pipeline templates

Save a working DeploymentPipeline as a versioned template, then instantiate it under a new name whenever a similar service ships.

Open guide →
Ops

Git webhooks

Map branches to pipelines so inbound GitHub, GitLab, or Bitbucket webhooks trigger the right run automatically, with signature verification.

Open guide →
Ops

Schedules

Attach a cron expression to any pipeline for recurring builds, deploys, or drift-detection runs — enable or disable without deleting.

Open guide →
Security

Policy engine

Built-in and Rego (OPA) checks block risky manifests — no :latest tags, no privileged containers, required labels — before GitOps sync.

Open guide →
Ops

Artifact registry

A built-in OCI registry for pipeline builds, with repository stats, tag counts, and total storage size at a glance.

Open guide →
Security

API keys

Issue org-scoped API keys for CI systems and automation with a one-time reveal, and revoke them the moment an integration retires.

Open guide →
Security

Audit log

Sensitive actions — pipeline changes, secret provider edits, key creation — land in a queryable, append-only audit trail.

Open guide →
Ops

Organization

Manage the organization your pipelines, agents, and keys belong to, and see exactly who has access.

GitOps

Apply manifests the way operators already think.

Target a Git repo, branch, and path. Tentacles clones, applies with server-side apply, waits for healthy, and can roll back when a deploy fails the gate.

  • Manifest sync

    Shallow clone and apply YAML from a target repo path — Argo-style without another control plane.

  • Health gates

    Deployments, StatefulSets, DaemonSets, and Pods must become ready before the step succeeds.

  • Policy & secrets

    Vault, AWS SM, GCP SM, and Kubernetes secrets resolve into the build and deploy steps.

Agents

Execution stays close to the cluster.

The control plane schedules work; agents on UAT, production, or edge environments do the building and applying — including promoting Officeless Visual Programming, Data Manager, and Workflows from Dev to PRD.

  • In-cluster or remote

    Run the agent next to workloads or on a bastion that holds kubeconfig for the target.

  • Least privilege

    Agents only get the RBAC and secret providers you attach — not the whole platform credential set.

  • Live execution log

    Stream build and sync output into the console so operators can approve or abort with context.

Security

Internal-first access on the VAS platform.

Tentacles runs on prod-vas next to Teknopus — TLS at the ingress, sign-in through Teknopus Identity, and secrets resolved only on the agent that needs them.

  • Teknopus Identity SSO (shared with Teknopus & FinOps)
  • Org-scoped API keys for automation
  • Audit log for sensitive actions
  • Secret providers stay out of the UI
  • Hosted on prod-vas with TLS via cert-manager

Built for platform teams

Use Tentacles for internal Mekari pipelines today. Expand to partner and bank environments through dedicated agents — not by exposing every kubeconfig to one UI.

Technology stack

Built on the same VAS platform primitives.

Tentacles sits beside Teknopus, FinOps, and Secrets — sharing Identity, GitOps, and the prod-vas cluster instead of inventing a parallel control plane.

  • Go control plane

    HTTP + gRPC API, pipeline engine, GitOps apply, and agent orchestration in one binary.

  • Next.js console

    Marketing site, feature guides, and the operator UI served from the same UI image.

  • Teknopus Identity

    Google sign-in through identity.vas.mekari.com — the same Identity stack as Teknopus, FinOps, and Secrets.

  • PostgreSQL + agents

    Durable control-plane state on prod-vas; least-privilege agents dial home over gRPC to build and apply.

How it works

Three steps from repo to healthy rollout.

  1. 1

    Define the pipeline

    Point Tentacles at your Dockerfile, registry, and GitOps manifest path.

    Read the guide →
  2. 2

    Attach an agent

    Register a cluster agent that can build images and apply to the target kube API.

    Read the guide →
  3. 3

    Ship with confidence

    Trigger from the console, webhook, or schedule — watch canary or blue/green land healthy.

    Read the guide →

Promoting Officeless apps instead? Start with the Visual Programming guide or browse every feature guide.

Ready to run your next deploy through Tentacles?

Open the console on VAS, register an agent against your UAT cluster, and ship an Officeless or data-manager pipeline end to end.

Ship builds and GitOps from one console.

Open console