presto-e2e-test

Automate end-to-end Presto testing on remote clusters with verifier, BEEST, goshadow, and shadow.

1|Updated Nov 29, 2021
One-click install
npx skills add https://github.com/mkarrmann/dotfiles --skill presto-e2e-test
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: presto-e2e-test
Source: https://github.com/mkarrmann/dotfiles/tree/main/agent_config/skills/presto-e2e-test
Command: npx skills add https://github.com/mkarrmann/dotfiles --skill presto-e2e-test

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

End-to-end Presto testing against remote clusters for correctness verification (verifier, BEEST, goshadow query replay) and performance regression testing (shadow A/B comparison, CPU analysis). It covers CLI spot-checks, query replay, querybank suites, and A/B performance benchmarking. NOT for local unit tests or checkstyle — see presto-build for those. Requires a deployed cluster — see presto-deploy.

Core Features & Use Cases

  • Integrates Verifier, BEEST, goshadow, and shadow tools to validate builds, replay queries, and benchmark performance on remote clusters.
  • Supports CLI spot-checks, query replay, and querybank suites to provide comprehensive coverage across distinct testing scenarios.
  • Enables safe remote testing workflows for end-to-end validation before production deployment, with guidance on cluster reservation and region considerations (presto-deploy).

Quick Start

Run presto-test verifier -c <cluster> --control <control_cluster> --suite <suite> to kick off a regression test on a reserved remote cluster.

Frequently Asked Questions about presto-e2e-test

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run end-to-end Presto tests against a remote cluster?

To run end-to-end Presto tests against a remote cluster, use the presto-test dispatcher to execute verifier, BEEST, goshadow, or shadow workflows on a reserved deployed cluster for correctness verification and performance regression analysis.

What is the difference between shadow A/B comparison and query replay for Presto testing?

Shadow A/B comparison benchmarks performance regressions between clusters, while query replay uses verifier, BEEST, and goshadow to validate correctness by replaying queries against control and target Presto clusters.

Do I need a deployed cluster to run Presto end-to-end correctness verification?

Yes, Presto end-to-end correctness verification requires a deployed remote cluster and the presto-test dispatcher along with supporting tooling like verifier, BEEST, goshadow, and shadow to execute workflows.

Can I run local unit tests or checkstyle with this Presto testing workflow?

No, this workflow does not handle local unit tests or checkstyle. It automates remote end-to-end Presto testing including CLI spot-checks, querybank suites, and A/B performance benchmarking; use presto-build for local checks.

How do I start a Presto verifier regression test on a reserved cluster?

Start a Presto verifier regression test by running presto-test verifier with the target cluster, control cluster, and suite parameters specified to execute query replay and validate correctness on the remote deployment.

When should I use querybank suites instead of CLI spot-checks for Presto testing?

Use querybank suites for comprehensive query replay coverage across multiple scenarios, whereas CLI spot-checks provide quick, targeted validation of specific Presto cluster behaviors before production deployment.