pumpwood-deploy-graph-datalake

Satellite deploy package for the Pumpwood Graph Datalake microservice on Kubernetes. It generates manifests for the API application, unified dataloader worker, and Graph Datalake secrets — then hands them to pumpwood-deploy for apply.

Developed by Murabei Data Science. BSD-3-Clause.

Pumpwood is a native Brazilian tree with a symbiotic relation to ants (Murabei)

---

Objective and motivation

This package turns Graph Datalake image tags into Kubernetes manifests that DeployPumpWood can apply in a Pumpwood cluster.

Why this exists

Graph Datalake lives in its own image repository (pumpwood-graph-datalake). Deploy scripts need a small, versioned Python package that knows the correct Secret, Deployment, and Service templates — without embedding YAML in every customer deploy repo.

How it is used

Project deploy scripts import PumpWoodGraphDatalakeMicroservice, pass image versions and DB credentials, register it with DeployPumpWood.add_microservice, then call deploy_microservices().

Scope

In scope: app + one unified dataloader worker + secrets. Out of scope: Postgres/PgBouncer, auth, storage ConfigMap, and the legacy split num/text edge workers.

---

What it deploys

Manifest	Kubernetes resources
pumpwood_graph_datalake__secrets	Secret pumpwood-graph-datalake
pumpwood_graph_datalake__deploy	Deployment + Service pumpwood-graph-datalake-app
pumpwood_graph_datalake_dataloader__worker	Deployment pumpwood-graph-datalake-worker-dataloader

Graph Datalake stores graph-oriented variables and edges. The app serves HTTP APIs; the dataloader worker consumes RabbitMQ messages and uploads data in parallel chunks.

flowchart LR
    subgraph pkg [pumpwood-deploy-graph-datalake]
        A[PumpWoodGraphDatalakeMicroservice]
    end
    subgraph core [pumpwood-deploy]
        B[DeployPumpWood]
    end
    subgraph cluster [Cluster]
        S[pumpwood-graph-datalake Secret]
        APP[pumpwood-graph-datalake-app]
        W[dataloader worker]
        RMQ[rabbitmq-main]
    end
    A --> B
    B --> S
    B --> APP
    B --> W
    RMQ --> APP
    RMQ --> W

---

Prerequisites

This package does not stand alone. Before Graph Datalake pods can start, the cluster must already provide:

Resource	Provided by
storage ConfigMap	StandardMicroservices in pumpwood-deploy
general-secrets	StandardMicroservices
rabbitmq-main-secrets	StandardMicroservices
Storage keys (GCP / Azure / AWS)	DeployPumpWood storage config
Postgres for Graph Datalake	PostgresDatabase + PGBouncerDatabase
Auth (typical)	pumpwood-deploy-auth

Storage bucket name and type are read from the cluster storage ConfigMap — they are not passed to PumpWoodGraphDatalakeMicroservice.

For local or CI dev databases, deploy Postgres/PgBouncer from pumpwood-deploy instead of an embedded test database.

---

Installation

pip install pumpwood-deploy-graph-datalake

Requires pumpwood-deploy.

---

Quick start

import os
import simplejson as json
from dotenv import load_dotenv
from pumpwood_deploy.deploy import DeployPumpWood
from pumpwood_deploy.microservices.postgres.deploy import (
    PostgresDatabase, PGBouncerDatabase)
from pumpwood_deploy_graph_datalake import (
    PumpWoodGraphDatalakeMicroservice)

with open("secrets/production.json", "r") as file:
    secrets = json.loads(file.read())
load_dotenv()

deploy = DeployPumpWood(
    model_user_password=secrets["microservices--model"],
    rabbitmq_secret=secrets["rabbitmq_secret"],
    hash_salt=secrets["hash_salt"],
    storage_type="aws_s3",
    storage_deploy_args={
        "storage_bucket_name": "my-pumpwood-bucket",
        "access_key_id": secrets["aws_access_key_id"],
        "secret_access_key": secrets["aws_secret_access_key"],
    },
    k8_provider="aws",
    k8_deploy_args={
        "region": "us-east-1",
        "cluster_name": "my-cluster",
    },
    k8_namespace="pumpwood",
)

deploy.add_microservice(
    PostgresDatabase(
        db_username="pumpwood",
        db_password=secrets["postgres_password"],
        name="postgres-main",
        disk_name="postgres-disk",
        disk_size="150Gi",
    ))

deploy.add_microservice(
    PGBouncerDatabase(
        name="pgbouncer-pumpwood-graph-datalake",
        postgres_database="pumpwood_graph_datalake",
        postgres_secret="postgres-main",
        postgres_host="postgres-main",
    ))

deploy.add_microservice(
    PumpWoodGraphDatalakeMicroservice(
        app_version=os.getenv("PUMPWOOD_GRAPH_DATALAKE_APP"),
        worker_version=os.getenv("PUMPWOOD_GRAPH_DATALAKE_WORKER"),
        repository="my-registry.example.com",
        db_host="pgbouncer-pumpwood-graph-datalake",
        db_database="pumpwood_graph_datalake",
        db_password=secrets["postgres_password"],
        microservice_password=secrets["microservice--graph-datalake"],
        app_replicas=1,
        app_debug="FALSE",
        worker_replicas=1,
        worker_n_parallel=4,
    ))

deploy.create_deploy_files()
deploy.deploy_microservices()

Environment variables

PUMPWOOD_GRAPH_DATALAKE_APP=1.0.0
PUMPWOOD_GRAPH_DATALAKE_WORKER=1.0.0

If the rendered manifest matches what is already on the cluster, kubectl apply produces no changes — safe for rolling image updates.

---

Configuration reference

Required

Parameter	Description
app_version	Image tag for pumpwood-graph-datalake-app
worker_version	Image tag for pumpwood-graph-datalake-worker-dataloader

Database

Parameter	Default	Description
db_host	pgbouncer-pumpwood-graph-datalake	Postgres host (use PgBouncer in prod)
db_port	5432	Postgres port
db_database	pumpwood	Database name
db_username	pumpwood	Database user
db_password	pumpwood	Database password
microservice_password	microservice--graph-datalake	Service user password
repository	GCR default	Docker registry for app and worker

Application

Parameter	Default	Description
app_replicas	1	Number of app pods
app_debug	FALSE	Debug flag
app_workers	10	Granian workers (GRANIAN_WORKERS)
app_timeout	300	Request timeout (seconds)
app_limits_memory	60Gi	Memory limit
app_limits_cpu	12000m	CPU limit
app_requests_memory	20Mi	Memory request
app_requests_cpu	1m	CPU request

Dataloader worker

Parameter	Default	Description
worker_replicas	1	Worker pod count
worker_debug	FALSE	Worker debug flag
worker_n_parallel	4	Parallel upload requests
worker_chunk_size	1000	Rows per parallel batch
worker_query_limit	1000000	Max rows per upload cycle
worker_limits_memory	60Gi	Worker memory limit
worker_limits_cpu	12000m	Worker CPU limit
worker_requests_memory	20Mi	Worker memory request
worker_requests_cpu	1m	Worker CPU request

---

Health check

The app Deployment exposes a readiness probe at:

GET /health-check/pumpwood-graph-datalake-app/  (port 5000)

Use this path for ingress and load balancer health checks.

---

Migration note

Older deploy scripts imported from the monolithic package:

# Before
from pumpwood_deploy.microservices.pumpwood_graph_datalake.deploy import (
    PumpWoodGraphDatalakeMicroservice)

# After
from pumpwood_deploy_graph_datalake import (
    PumpWoodGraphDatalakeMicroservice)

Removed from the satellite API (use cluster-level config instead):

• bucket_name — now from storage ConfigMap
• test_db_* / embedded Postgres — use PostgresDatabase / PGBouncerDatabase from core
• worker_num_version / worker_text_version — replaced by a single worker_version for the unified dataloader image

The monolithic package deployed five manifests (optional Postgres + two edge workers). This satellite deploys three (app + unified worker + secrets).

---

Related packages

Package	Role
pumpwood-deploy	Orchestrator, Kong, RabbitMQ, Postgres
pumpwood-deploy-datalake	Standard datalake microservice
pumpwood-deploy-auth	Authorization microservice

Full platform documentation: Murabei Open Source — pumpwood-deploy.

---

Development

pip install -e ../pumpwood-deploy
pip install -e .

PYTHONPATH="src:../pumpwood-deploy/src" \
  python3 -m unittest discover \
  -s src/pumpwood_deploy_graph_datalake/tests -p "test_*.py" -v

ruff check src/

Bump VERSION via ./build.sh before publishing (same flow as sibling satellite deploy packages).