Happy Infra Metrics and Grafana

What it does

When your agent uses it

• Querying current or historical Prometheus metrics for Happy's infrastructure
• Pulling and editing a Grafana dashboard as code
• Adding or modifying a dashboard panel

Example prompts

• “Pull the current state of the API latency dashboard so I can edit it.”
• “Run a range query for request rate over the last hour through Grafana's proxy.”
• “Push my updated dashboard JSON without locking it from UI edits.”

Requirements

• grafanactl (installed via Go)
• Grafana and Prometheus credentials in a local .env file

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

1. 1Use the Prometheus datasource: {"type": "prometheus", "uid": "$GRAFANA_PROMETHEUS_UID"}
2. 2Pick the next available id (check existing panels for max id)
3. 3Position with gridPos: h = height (8 standard), w = width (12 half, 24 full), x = column (0 or 12), y = row
4. 4Common panel types: stat, timeseries, piechart, bargauge, table
5. 5Set appropriate unit: percentunit, ops, s, short, reqps

What it can do on your machine

Read from SKILL.md and the folder at commit fba320e. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Shell commands in SKILL.md call:

  • curl
  • python3
  • go

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md. Its commands use curl, which can reach the network depending on how they are called.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names these keys or tokens, usually read from environment variables:

  • GRAFANA_PASSWORD

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Happy Infra Metrics and Grafana loads about 2k tokens when it runs. Until then it costs about 81 tokens; SKILL.md has 422 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~81

When it runs · the whole SKILL.md, loaded when a task matches

~2k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

• NoteMentions a .env fileSKILL.md:16

  Credentials are stored in the repo root `.env` file (gitignored). Load them before running commands:

• NoteMentions a .env fileSKILL.md:27

  set -a; source .env; set +a

• NoteMentions a .env fileSKILL.md:48

  set -a; source .env; set +a

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from slopus/happy at commit fba320e, republished under its MIT licence (© slopus). 422 words, ~1,985 tokens.

name

metrics-graphana

description

Query and manage Grafana dashboards and Prometheus metrics for Happy infrastructure. Covers grafanactl CLI usage, direct Prometheus queries through Grafana proxy, and dashboard-as-code workflows. Use when user asks about metrics, dashboards, monitoring, Grafana, Prometheus, or wants to add/modify panels.

GRAFANA_URL=...
GRAFANA_USER=...
GRAFANA_PASSWORD=...
GRAFANA_PROMETHEUS_UID=...

set -a; source .env; set +a

go install github.com/grafana/grafanactl/cmd/grafanactl@latest

# Load env vars first
set -a; source .env; set +a

# Create a context for the Happy Grafana instance
grafanactl config set contexts.happy.grafana.server "$GRAFANA_URL"
grafanactl config set contexts.happy.grafana.user "$GRAFANA_USER"
grafanactl config set contexts.happy.grafana.password "$GRAFANA_PASSWORD"
grafanactl config set contexts.happy.grafana.org-id 1

# Switch to the context
grafanactl config use-context happy

# Verify
grafanactl config check

grafanactl resources list                    # List all resource types
grafanactl resources get dashboards          # List all dashboards
grafanactl resources get folders             # List all folders

grafanactl resources pull dashboards -p ./resources -o json
grafanactl resources pull dashboards/DASHBOARD_ID -p ./resources -o json

# Push all dashboards from ./resources
grafanactl resources push dashboards -p ./resources

# Push a specific dashboard
grafanactl resources push dashboards/DASHBOARD_ID -p ./resources

# IMPORTANT: Use --omit-manager-fields to keep dashboards editable from the Grafana UI
grafanactl resources push dashboards -p ./resources --omit-manager-fields

# Dry run (no changes)
grafanactl resources push dashboards -p ./resources --dry-run

# 1. Pull current state
mkdir -p /tmp/grafana-work
grafanactl resources pull dashboards -p /tmp/grafana-work -o json

# 2. Edit the JSON files (add panels, modify queries, etc.)

# 3. Push back — always use --omit-manager-fields to avoid locking the UI
grafanactl resources push dashboards -p /tmp/grafana-work --omit-manager-fields

curl -s -u "$GRAFANA_USER:$GRAFANA_PASSWORD" \
  --data-urlencode 'query=YOUR_PROMQL_HERE' \
  "$GRAFANA_URL/api/datasources/proxy/uid/$GRAFANA_PROMETHEUS_UID/api/v1/query" \
  | python3 -m json.tool

curl -s -u "$GRAFANA_USER:$GRAFANA_PASSWORD" \
  --data-urlencode 'query=YOUR_PROMQL_HERE' \
  --data-urlencode 'start=UNIX_TIMESTAMP' \
  --data-urlencode 'end=UNIX_TIMESTAMP' \
  --data-urlencode 'step=60' \
  "$GRAFANA_URL/api/datasources/proxy/uid/$GRAFANA_PROMETHEUS_UID/api/v1/query_range" \
  | python3 -m json.tool

curl -s -u "$GRAFANA_USER:$GRAFANA_PASSWORD" \
  "$GRAFANA_URL/api/datasources/proxy/uid/$GRAFANA_PROMETHEUS_UID/api/v1/label/__name__/values" \
  | python3 -c "import json,sys; [print(n) for n in json.load(sys.stdin)['data']]"

# Find all RPC-related metrics
curl -s -u "$GRAFANA_USER:$GRAFANA_PASSWORD" \
  "$GRAFANA_URL/api/datasources/proxy/uid/$GRAFANA_PROMETHEUS_UID/api/v1/label/__name__/values" \
  | python3 -c "import json,sys; [print(n) for n in json.load(sys.stdin)['data'] if 'rpc' in n.lower()]"

# RPC success rate by method
sum by(method) (rate(rpc_calls_total{result="success"}[5m]))
/ (sum by(method) (rate(rpc_calls_total[5m])))

# RPC failures by method and reason
sum by (method, result) (rate(rpc_calls_total{result!="success"}[5m]))

# RPC failures by type only
sum by (result) (rate(rpc_calls_total{result!="success"}[5m]))

# RPC P95 latency by method
histogram_quantile(0.95, sum by (method, le) (rate(rpc_call_duration_seconds_bucket[5m])))

# Socket lookup retry distribution (P95)
histogram_quantile(0.95, sum by (method, le) (rate(rpc_lookup_retries_bucket[5m])))

# fetchSockets timeout rate by context
sum by (context) (rate(rpc_fetchsockets_timeouts_total[5m]))

# HTTP error rate
sum(rate(http_requests_total{status=~"5.."}[5m])) / sum(rate(http_requests_total[5m]))

# Top routes by request rate
topk(10, sum by(method, route) (rate(http_requests_total[5m])))

© slopus, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .agents/skills/metrics-graphana of slopus/happy.

Open the folder on GitHubat commit fba320e