playbooks

home / skills / tdhopper / dotfiles2.0 / monitor-deploy

monitor-deploy skill

/.claude/skills/monitor-deploy

This skill helps you monitor Spotify deployments by using Tugboat CLI, showing status, progress, and optionally tails logs.

npx playbooks add skill tdhopper/dotfiles2.0 --skill monitor-deploy

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
1.5 KB
---
name: monitoring-deployments
description: Use this skill when monitoring Spotify deployments, checking deployment status, viewing deploy progress, or tailing deployment logs. Triggers on "check deployment", "deploy status", "monitor deploy", "what's deploying", or when user runs /monitor-deploy.
---

# Monitoring Deployments

Monitor Spotify deployments via the Tugboat CLI with optional GCP log tailing.

## Prerequisites

Check if `tugboat` exists. If not, auto-install:

```bash
uv tool install --from git+ssh://[email protected]/warpspeed/tugboat-cli.git tugboat-cli
```

## Workflow

### 1. Get component

Use provided component name/pattern, or ask the user.

### 2. List installations

```bash
tugboat installations list --component=<component-or-pattern>
```

### 3. Get deployment status

```bash
tugboat deployments show --installation=<installation-id> <deployment-id>
```

For full JSON details:

```bash
tugboat deployments dump --installation=<installation-id> <deployment-id>
```

### 4. Get version info

```bash
tugboat version <component-id> show <version-id>
```

### 5. Summarize

Present concisely: state, progress %, version/commit, errors.

### 6. Offer log tailing

When in-progress or failed, offer to tail GCP logs. Parse log info from tugboat output and run:

```bash
gcloud logging tail "<filter>" --project=<project>
```

## Batch mode

For patterns like `fan-audio-*`, show a table of all matching components.

## Errors

- Auth issues → `gcloud auth login`
- Missing tugboat → auto-install
- Bad component → list available ones

Overview

This skill monitors Spotify deployments using the Tugboat CLI and can optionally tail GCP logs for live insight. It helps check deployment status, view progress, summarize version and errors, and follow logs when a deployment is in progress or failed. Use it interactively or via the /monitor-deploy trigger for quick checks.

How this skill works

The skill ensures tugboat is available and auto-installs it if missing. It lists installations for a component, fetches deployment details and version information, and summarizes state, progress, version/commit, and errors. When needed, it parses log filters from tugboat output and runs gcloud logging tail to stream logs for the relevant project.

When to use it

  • Check current deployment status for a specific component or installation
  • Monitor progress of an ongoing deployment and estimate completion
  • Investigate failed deployments and surface errors and failing steps
  • Tail production or staging logs when a deployment is stuck or exhibiting issues
  • Quickly scan multiple components with a name pattern in batch mode

Best practices

  • Provide a precise component name or pattern to avoid long lists
  • Use installation IDs when available for direct lookups
  • If tugboat is missing, allow the tool to auto-install to avoid interruptions
  • When tailing logs, ensure you are authenticated with gcloud (gcloud auth login)
  • Use the full JSON dump when you need detailed fields or trace IDs for log correlation

Example use cases

  • Check deployment status: run tugboat deployments show to get state, progress, and errors
  • Batch monitor: query installations for fan-audio-* and display a table of matching deployments
  • Investigate failure: fetch deployments dump for detailed JSON, then tail GCP logs filtered by the parsed filter
  • Confirm version: run tugboat version <component-id> show <version-id> to validate the deployed commit
  • Live troubleshooting: when a deployment is stuck, stream gcloud logging tail with the parsed filter to watch events in real time

FAQ

What if tugboat is not installed?

The skill auto-installs tugboat using the configured tool installer command. If that fails, install tugboat manually with the provided uv tool install command.

How do I stream logs for a failing deployment?

When the deployment is in-progress or failed, the skill extracts a log filter from tugboat output and runs gcloud logging tail "<filter>" --project=<project>. Make sure you are authenticated with gcloud (gcloud auth login).