Skip to content

README.md

Latest commit

 

History

History
134 lines (97 loc) · 4.32 KB

README.md

File metadata and controls

134 lines (97 loc) · 4.32 KB

lkr: Looker Load Testing & Embed User Management CLI

A CLI utility built using Typer, Locust, and the Looker Python SDK to execute load tests, collect embed observability metrics, and manage/cleanup Looker SSO embed users.

Features

  • Debug Diagnostics: Quickly test and verify your Looker API connection and SDK settings.
  • Load Testing: Simulates concurrent users accessing dashboards, running queries, rendering PDFs/images, or loading cookieless embed environments.
  • Observability Server: Hosts a local dashboard iframe server, launches automated users, tracks Looker iframe JS event timings (loaded, run:start, run:complete), and reports performance metrics.
  • Embed User Cleanup: Batch deletes expired/old SSO embed users using concurrent thread pools to keep Looker instance user lists clean.

Getting Started

Prerequisites

  • Python 3.13 or higher
  • uv (recommended Python package installer)
  • Docker (optional, for containerization)
  • Google Cloud SDK (gcloud) (optional, if pushing to GCP Artifact Registry)

Installation

Clone/copy the repository, navigate to the folder, and run:

# Install runtime dependencies
make install

# Install dev dependencies (pytest, pre-commit, etc.)
make dev-install

Configuration

The CLI relies on standard Looker SDK environment variables. Create a .env file in the root of the project:

LOOKERSDK_BASE_URL="https://your-looker-instance.com:19999"
LOOKERSDK_CLIENT_ID="your_api_client_id"
LOOKERSDK_CLIENT_SECRET="your_api_client_secret"

Verify your configuration by running the debug check:

make run-debug

Command Usage Reference

The main entry point is lkr.

1. Verification

# Debug credentials & check connectivity
uv run lkr load-test debug looker

2. Load Testing Dashboards

# Run a dashboard load test for 5 minutes with 25 users
uv run lkr load-test dashboard --dashboard <dashboard_id> --model <model_name> --users 25 --run-time 5

3. Load Testing Cookieless Embed Dashboards

uv run lkr load-test cookieless-embed-dashboard --dashboard <dashboard_id> --model <model_name> --users 25

4. Load Testing Queries (Explore QIDs)

uv run lkr load-test query --query <query_id> --model <model_name> --users 25

5. Load Testing Dashboard Rendering (PDF/PNG/JPG)

uv run lkr load-test render --dashboard <dashboard_id> --model <model_name> --result-format pdf

6. Embed Observability Load Testing

Starts a local web server to capture Looker javascript frontend metrics during simulated dashboard sessions:

uv run lkr load-test embed-observability --dashboard <dashboard_id> --port 4000

7. Deleting Embed Users (Cleanup)

SSO embed logins create new users in Looker. Clean them up using concurrent batch deletion:

# Dry run to see how many users match the prefix (defaults to first name "Embed")
uv run lkr load-test delete-embed-users --first-name "Embed" --dry-run

# Execute deletion with a limit
uv run lkr load-test delete-embed-users --first-name "Embed" --no-dry-run --limit 500

Deployment

This tool is pre-packaged as a Docker image. You do not need to install Python or build any code to use it. You can run it effortlessly on a serverless Google Cloud Run Job by copying the command below:

  1. Update --project to your Google Cloud project.
  2. Provide your Looker API Credentials in the --set-env-vars flag.
  3. Pass any script arguments into --args.
gcloud run jobs create delete-embed-users-job \
    --project=your-target-cloud-project \
    --region=us-central1 \
    --task-timeout=60 \
    --max-retries=0 \
    --execute-now \
    --image=us-west1-docker.pkg.dev/looker-scale-testing/delete-embed-users-v2/delete-embed-users:latest \
    --set-env-vars=LOOKERSDK_CLIENT_ID=YOUR_CLIENT_ID,LOOKERSDK_CLIENT_SECRET=YOUR_CLIENT_SECRET,LOOKERSDK_BASE_URL=https://yourinstance.cloud.looker.com \
    --command="lkr" \
    --args="delete-embed-users"

Development

Use the included Makefile for common development tasks:

  • make test - Run pytest suites.
  • make format - Format code and run pre-commit checks.
  • make clean - Remove virtualenv, caches, and build artifacts.