How do I install the Digitalocean App Insights skill?

Run npx skillmds@latest add willakins/digitalocean-app-insights in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Digitalocean App Insights safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Digitalocean App Insights?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Digitalocean App Insights free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Digitalocean App Insights?

willakins (@willakins) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Digitalocean App Insights by willakins

Name: Digitalocean App Insights
Author: willakins

Digitalocean App Insights

Investigate DigitalOcean App Platform health, deployments, instance sizing, CPU/memory/restart metrics, and memory pressure using the DigitalOcean API. Use when Codex needs to inspect a DigitalOcean app, diagnose production app resource pressure, correlate Sentry or app symptoms with App Platform metrics, or ask the user for the minimum read-only DigitalOcean token needed for live observability.

willakins Updated Sep 17, 2026

File contents

DigitalOcean App Insights

Use this skill for read-only DigitalOcean App Platform observability. Prefer the bundled helper, keep credential handling local, and separate observed metrics from inferred causes.

Read API notes for endpoint details, scopes, response shapes, Kubernetes follow-up, and output fields.

Safety Model

Use read-only API access by default. Do not ask the user to paste tokens in chat. Ask them to set DIGITALOCEAN_ACCESS_TOKEN locally and confirm; accept DIGITALOCEAN_TOKEN as a fallback. Never print tokens or credential-bearing URLs.

If the token is missing or under-scoped, stop and ask for the smallest needed read scope. Do not request write scopes.

Quick Start

Prefer the bundled helper:

python3 ~/.codex/skills/digitalocean-app-insights/scripts/do_app_insights.py list-apps
python3 ~/.codex/skills/digitalocean-app-insights/scripts/do_app_insights.py inspect --app-name croft-prod --hours 6
python3 ~/.codex/skills/digitalocean-app-insights/scripts/do_app_insights.py inspect --app-id APP_UUID --component web --hours 24

Use --json when raw payloads are needed for deeper analysis. Use --base-url only for testing.

Investigation Workflow

Resolve the target app by id or exact/partial name. Ask when multiple apps match.
Record service names, instance count, size slug, autoscaling settings, active deployment, and recent deployments.
Pull recent memory_percentage, cpu_percentage, and restart_count; start with 6h for incidents, 24h for slow-burn memory pressure, and 7d for trends.
Inspect app health and instances to spot one hot instance or component.
Correlate timestamps with Sentry events, deploys, logs, or database metrics.

Interpretation

Treat thresholds as heuristics. >= 90% memory is immediate pressure or OOM risk; sustained 80-90% is capacity risk. High memory plus restart increases is consistent with OOM/restart pressure; high memory without restarts may be cache growth, workload shape, small instance size, or normal high watermark.

Use Kubernetes access only when the user explicitly asks and provides access; keep commands read-only.

---name: digitalocean-app-insightsdescription: Investigate DigitalOcean App Platform health, deployments, instance sizing, CPU/memory/restart metrics, and memory pressure using the DigitalOcean API. Use when Codex needs to inspect a DigitalOcean app, diagnose production app resource pressure, correlate Sentry or app symptoms with App Platform metrics, or ask the user for the minimum read-only DigitalOcean token needed for live observability.---# DigitalOcean App InsightsUse this skill for read-only DigitalOcean App Platform observability. Prefer the bundled helper, keep credential handling local, and separate observed metrics from inferred causes.Read [API notes](references/api-notes.md) for endpoint details, scopes, response shapes, Kubernetes follow-up, and output fields.## Safety ModelUse read-only API access by default. Do not ask the user to paste tokens in chat. Ask them to set `DIGITALOCEAN_ACCESS_TOKEN` locally and confirm; accept `DIGITALOCEAN_TOKEN` as a fallback. Never print tokens or credential-bearing URLs.If the token is missing or under-scoped, stop and ask for the smallest needed read scope. Do not request write scopes.## Quick StartPrefer the bundled helper:```bashpython3 ~/.codex/skills/digitalocean-app-insights/scripts/do_app_insights.py list-appspython3 ~/.codex/skills/digitalocean-app-insights/scripts/do_app_insights.py inspect --app-name croft-prod --hours 6python3 ~/.codex/skills/digitalocean-app-insights/scripts/do_app_insights.py inspect --app-id APP_UUID --component web --hours 24```Use `--json` when raw payloads are needed for deeper analysis. Use `--base-url` only for testing.## Investigation Workflow1. Resolve the target app by id or exact/partial name. Ask when multiple apps match.2. Record service names, instance count, size slug, autoscaling settings, active deployment, and recent deployments.3. Pull recent `memory_percentage`, `cpu_percentage`, and `restart_count`; start with 6h for incidents, 24h for slow-burn memory pressure, and 7d for trends.4. Inspect app health and instances to spot one hot instance or component.5. Correlate timestamps with Sentry events, deploys, logs, or database metrics.## InterpretationTreat thresholds as heuristics. `>= 90%` memory is immediate pressure or OOM risk; sustained `80-90%` is capacity risk. High memory plus restart increases is consistent with OOM/restart pressure; high memory without restarts may be cache growth, workload shape, small instance size, or normal high watermark.Use Kubernetes access only when the user explicitly asks and provides access; keep commands read-only.

Frequently asked questions