Managing Looker

Looker Management Skill

Manage and analyze Looker explores, dashboards, LookML projects, and query performance via the Looker API.

MANDATORY: Discovery-First Pattern

Always discover projects and models before querying specific explores or dashboards.

Phase 1: Discovery

#!/bin/bash

looker_api() {
    local method="${1:-GET}"
    local endpoint="$2"
    local data="${3:-}"

    if [ -n "$data" ]; then
        curl -s -X "$method" \
            -H "Authorization: token ${LOOKER_API_TOKEN}" \
            -H "Content-Type: application/json" \
            "${LOOKER_BASE_URL}/api/4.0/${endpoint}" \
            -d "$data"
    else
        curl -s -X "$method" \
            -H "Authorization: token ${LOOKER_API_TOKEN}" \
            "${LOOKER_BASE_URL}/api/4.0/${endpoint}"
    fi
}

echo "=== LookML Projects ==="
looker_api GET "projects" | jq -r '
    .[] | "\(.id)\t\(.name)\t\(.git_service_name // "N/A")\t\(if .is_example then "EXAMPLE" else "CUSTOM" end)"
' | column -t

echo ""
echo "=== LookML Models ==="
looker_api GET "lookml_models" | jq -r '
    .[] | "\(.name)\t\(.project_name)\t\(.explores | length) explores"
' | column -t | head -20

echo ""
echo "=== Dashboards (recent) ==="
looker_api GET "dashboards?fields=id,title,folder,view_count&sorts=view_count desc&limit=20" | jq -r '
    .[] | "\(.id)\t\(.title[0:40])\t\(.folder?.name // "?")\t\(.view_count) views"
' | column -t

Core Helper Functions

#!/bin/bash

looker_api() {
    local method="${1:-GET}"
    local endpoint="$2"
    local data="${3:-}"

    if [ -n "$data" ]; then
        curl -s -X "$method" \
            -H "Authorization: token ${LOOKER_API_TOKEN}" \
            -H "Content-Type: application/json" \
            "${LOOKER_BASE_URL}/api/4.0/${endpoint}" \
            -d "$data"
    else
        curl -s -X "$method" \
            -H "Authorization: token ${LOOKER_API_TOKEN}" \
            "${LOOKER_BASE_URL}/api/4.0/${endpoint}"
    fi
}

Output Rules

• TOKEN EFFICIENCY: Target ≤50 lines per output
• Use fields query parameter to limit returned fields from Looker API
• Never dump full LookML model definitions — extract key metadata only

Common Operations

Explore Analysis

#!/bin/bash
MODEL="${1:?Model name required}"

echo "=== Explores in Model: $MODEL ==="
looker_api GET "lookml_models/${MODEL}" | jq -r '
    .explores[] | "\(.name)\t\(.label // .name)\t\(.description[0:50] // "no description")"
' | column -t | head -20

echo ""
echo "=== Explore Fields ==="
EXPLORE="${2:-}"
if [ -n "$EXPLORE" ]; then
    looker_api GET "lookml_models/${MODEL}/explores/${EXPLORE}?fields=fields" | jq '{
        dimensions: (.fields.dimensions | length),
        measures: (.fields.measures | length),
        filters: (.fields.filters | length),
        parameters: (.fields.parameters | length)
    }'

    echo ""
    echo "=== Top Dimensions ==="
    looker_api GET "lookml_models/${MODEL}/explores/${EXPLORE}?fields=fields" | jq -r '
        .fields.dimensions[:10][] | "\(.name)\t\(.type)\t\(.label_short // .label)"
    ' | column -t
fi

Dashboard Analysis

#!/bin/bash
DASHBOARD_ID="${1:?Dashboard ID required}"

echo "=== Dashboard Details ==="
looker_api GET "dashboards/${DASHBOARD_ID}" | jq '{
    id: .id,
    title: .title,
    folder: .folder?.name,
    elements: (.dashboard_elements | length),
    filters: (.dashboard_filters | length),
    view_count: .view_count,
    last_viewed_at: .last_viewed_at
}'

echo ""
echo "=== Dashboard Elements ==="
looker_api GET "dashboards/${DASHBOARD_ID}" | jq -r '
    .dashboard_elements[] | "\(.id)\t\(.type)\t\(.title // "untitled")\t\(.look_id // .query_id // "?")"
' | column -t | head -20

echo ""
echo "=== Dashboard Filters ==="
looker_api GET "dashboards/${DASHBOARD_ID}" | jq -r '
    .dashboard_filters[] | "\(.name)\t\(.type)\t\(.default_value // "none")\t\(.model)\t\(.explore)"
' | column -t

PDT Status Monitoring

#!/bin/bash
echo "=== PDT Builds ==="
looker_api GET "derived_table/graph/model" 2>/dev/null | jq -r '
    to_entries[] | "\(.key)\t\(.value.status // "?")\t\(.value.last_build_started_at[0:16] // "never")"
' | column -t | head -20

echo ""
echo "=== Running PDT Builds ==="
looker_api GET "running_queries?fields=id,status,source,model,explore,created_at" | jq -r '
    .[] | select(.source == "pdt") |
    "\(.id)\t\(.status)\t\(.model)\t\(.explore)\t\(.created_at[0:16])"
' | column -t

echo ""
echo "=== PDT Connection Status ==="
looker_api GET "connections" | jq -r '
    .[] | "\(.name)\t\(.pdt_enabled)\t\(.dialect_name)\t\(.host // "?")"
' | column -t | head -10

Query Performance

#!/bin/bash
echo "=== Slow Running Queries ==="
looker_api GET "running_queries?fields=id,user,query,runtime,status,source,created_at&limit=20" | jq -r '
    .[] | select(.runtime > 30) |
    "\(.id)\t\(.runtime | floor)s\t\(.status)\t\(.source)\t\(.user?.display_name // "system")"
' | column -t | head -15

echo ""
echo "=== Query History (by model) ==="
looker_api GET "query_tasks?fields=id,status,result_source,runtime" 2>/dev/null | jq '
    group_by(.result_source) | map({source: .[0].result_source, count: length, avg_runtime: ([.[].runtime] | add / length | floor)}) |
    .[] | "\(.source)\t\(.count) queries\tavg \(.avg_runtime)s"
' -r 2>/dev/null

echo ""
echo "=== Kill Long-Running Query ==="
echo "Use: looker_api DELETE 'running_queries/{query_task_id}'"

LookML Project Validation

#!/bin/bash
PROJECT="${1:?Project name required}"

echo "=== LookML Validation ==="
looker_api POST "projects/${PROJECT}/lookml_validation" | jq '{
    stale_content: .stale_content,
    errors: [.errors[] | {severity: .severity, kind: .kind, message: .message[0:80]}] | length,
    warnings: [.warnings[]? | .message[0:60]] | length
}'

echo ""
echo "=== Validation Errors ==="
looker_api GET "projects/${PROJECT}/lookml_validation" | jq -r '
    .errors[:10][] | "\(.severity)\t\(.kind)\t\(.message[0:80])"
' | column -t

echo ""
echo "=== Git Status ==="
looker_api GET "projects/${PROJECT}/git/branch" | jq '{
    branch: .name,
    ref: .ref,
    is_production: .is_production,
    ahead: .ahead_count,
    behind: .behind_count
}'

Output Format

Present results as a structured report:

Managing Looker Report
══════════════════════
Resources discovered: [count]

Resource       Status    Key Metric    Issues
──────────────────────────────────────────────
[name]         [ok/warn] [value]       [findings]

Summary: [total] resources | [ok] healthy | [warn] warnings | [crit] critical
Action Items: [list of prioritized findings]

Target ≤50 lines of output. Use tables for multi-resource comparisons.

Anti-Hallucination Rules

1. NEVER assume resource names — always discover via CLI/API in Phase 1 before referencing in Phase 2.
2. NEVER fabricate metric names or dimensions — verify against the service documentation or --help output.
3. NEVER mix CLI commands between service versions — confirm which version/API you are targeting.
4. ALWAYS use the discovery → verify → analyze chain — every resource referenced must have been discovered first.
5. ALWAYS handle empty results gracefully — an empty response is valid data, not an error to retry.

Counter-Rationalizations

Common Pitfalls

• API versions: Looker API 4.0 is current — older instances may only support 3.1 — check /api/4.0/versions
• Token auth: API tokens expire — if getting 401, re-authenticate via POST /api/4.0/login with client credentials
• PDT builds: PDTs build on schedule or trigger — stale PDTs mean stale dashboard data but no visible error
• Content validation: LookML validation errors don't prevent deployment in dev mode — production deploy may break
• Explore permissions: Users may not see all explores — API results depend on the authenticating user's permissions
• Folder structure: Dashboards live in folders with access controls — personal folders are per-user
• Dashboard vs Look: Dashboards contain multiple elements; Looks are single saved queries — they have different APIs
• Query caching: Looker caches query results — stale data may be from cache, not a query failure
• Model sets: Permissions are tied to model sets — a user may access some models but not others