Kubernetes Manifests & Workload Scripts - #6895
Conversation
5aa0a71 to
57af39a
Compare
| @@ -0,0 +1,64 @@ | |||
| # Vibe Coding Startup Scripts | |||
|
|
|||
| Pluggable startup scripts for the UC-A snapshot saturation harness (`sweeps/snapshot_saturation_search.py`). Each script simulates a realistic "vibe coding" sandbox cold-start — the kind of environment setup that happens when an AI coding agent provisions a new sandbox for a user. | |||
| @@ -0,0 +1,64 @@ | |||
| # Vibe Coding Startup Scripts | |||
There was a problem hiding this comment.
Let's drop "Vibe Coding" from the title and code path
There was a problem hiding this comment.
Done. I scrubbed "Vibe Coding" everywherem and also renamed the directory from vibe_coding/ to startup_scripts/.
| @@ -0,0 +1,84 @@ | |||
| #!/bin/bash | |||
| # Vibe Coding Startup Script — npm + Vite dev server | |||
There was a problem hiding this comment.
Let's drop "Vibe Coding" from the title and code path
There was a problem hiding this comment.
Done. I scrubbed "Vibe Coding" everywherem and also renamed the directory from vibe_coding/ to startup_scripts/.
|
|
||
| ## How It Works | ||
|
|
||
| When `--preload_mode=script:<path>` is passed to the sweep harness: |
There was a problem hiding this comment.
It's unclear to me whether both startup_pip_fastapi and startup_npm_vite scripts are used by the benchmarks. Can you clarify? Is it true that some benchmarks will use startup_pip_fastapi while the other will use startup_npm_vite?
There was a problem hiding this comment.
They are not run simultaneously. The user selects one of these scripts via the --k8s_snapshot_preload_mode CLI flag to test different workload footprints. Updated the README to clarify this.
|
|
||
| ### startup_pip_fastapi.sh | ||
|
|
||
| **Lightweight Python variant.** Runs natively in the `python:3.11-slim` base image. |
There was a problem hiding this comment.
Can you elaborate what these startup scripts do in the documentation?
There was a problem hiding this comment.
yes, updated the startup_scripts/README.md.
| | |-- snapshot-crds.yaml.j2 # PodSnapshot StorageConfig + Policy | ||
| | +-- snapshot-sandbox-template.yaml.j2 # Snapshot-specific SandboxTemplate | ||
| +-- workloads/ # Benchmark workload scripts | ||
| +-- vibe_coding/ # Startup scripts for snapshot benchmarks |
There was a problem hiding this comment.
Startup scripts for snapshot benchmarks
Please elaborate. Are these startup scripts only used by the snapshot benchmark?
There was a problem hiding this comment.
Correct, added more elaboration in the added README. Basically for the "vibe coding" scenario we discussed earlier; pip_fastapi.sh is the "Lightweight Python" workload (typically 5-8s cold start), and npm_vite.sh is the "Heavier Node.js" workload (typically 30-60s cold start).
| +-- vibe_coding/ # Startup scripts for snapshot benchmarks | ||
| |-- README.md | ||
| |-- startup_pip_fastapi.sh | ||
| +-- startup_npm_vite.sh |
There was a problem hiding this comment.
Please document the characteristics of these startup scripts.
There was a problem hiding this comment.
Done. added a new startup_scripts/README.md file.
2 participants
Files: ~10 new files (no Python benchmark logic)
perfkitbenchmarker/data/k8s_agents/README.mdperfkitbenchmarker/data/k8s_agents/manifests/adk-agent.yaml.j2perfkitbenchmarker/data/k8s_agents/manifests/sandbox-templates.yaml.j2perfkitbenchmarker/data/k8s_agents/manifests/sandbox-router.yaml.j2perfkitbenchmarker/data/k8s_agents/manifests/psi-reader.yaml.j2perfkitbenchmarker/data/k8s_agents/manifests/snapshot-crds.yaml.j2perfkitbenchmarker/data/k8s_agents/manifests/snapshot-sandbox-template.yaml.j2perfkitbenchmarker/data/k8s_agents/workloads/startup_scripts/npm_vite.shperfkitbenchmarker/data/k8s_agents/workloads/startup_scripts/pip_fastapi.shperfkitbenchmarker/data/k8s_agents/workloads/startup_scripts/README.mdDescription: Adds Jinja2 manifest templates for deploying the Agent Sandbox ecosystem onto GKE, plus vibe coding startup scripts for snapshot benchmarks. Includes: