Index Health & Analytics

Monitor your documentation index health and search performance with comprehensive analytics. Get real-time visibility into coverage, quality, latency, and usage patterns.

What You Get

Health Score — Composite 0-100 score based on coverage, quality, and freshness
Latency Percentiles — p50, p75, p90, p95, p99 response time metrics
Tool Usage Breakdown — See which MCP tools are most used
Error Analysis — Categorized error breakdown with examples
Recommendations — Actionable suggestions to improve index health

Index Health Monitoring

The Index Health panel provides a real-time view of your documentation index status. The health score is calculated from three factors:

Factor	Weight	Description
Coverage	40%	Percentage of documents with indexed chunks
Quality	30%	Average chunk quality score (structure, content)
Freshness	30%	Percentage of documents updated within threshold

Health Status Levels

80-100 Healthy

50-79 Warning

0-49 Critical

Tier Distribution

Chunks are categorized by access recency:

HOT

Recently accessed

WARM

Default tier

COLD

Rarely used

Search Analytics

The Search Analytics panel tracks query performance over configurable time periods (7, 14, 30, or 90 days).

Latency Percentiles

Understand your response time distribution:

Percentile	Description	Use Case
p50	Median response time	Typical user experience
p90	90th percentile	Most users experience
p95	95th percentile	SLA monitoring
p99	99th percentile	Tail latency

Error Categories

Errors are automatically categorized for easy analysis:

rate_limit

Too many requests (429)

timeout

Query exceeded time limit

auth_error

Authentication/authorization failed

validation_error

Invalid input parameters

MCP Tools

Access analytics programmatically via MCP tools:

rlm_index_health

result = mcp_client.call_tool("rlm_index_health", {
    "stale_threshold_days": 30
})

print(f"Health: {result['health_score']}/100")
print(f"Coverage: {result['coverage_percent']}%")
print(f"Stale docs: {result['stale_count']}")

rlm_search_analytics

analytics = mcp_client.call_tool("rlm_search_analytics", {
    "days": 30
})

print(f"Queries: {analytics['summary']['total_queries']}")
print(f"Success rate: {analytics['summary']['success_rate'] * 100}%")
print(f"p95 latency: {analytics['latency']['p95']}ms")

rlm_index_recommendations

recs = mcp_client.call_tool("rlm_index_recommendations", {})

for rec in recs["recommendations"]:
    print(f"[{rec['severity']}] {rec['title']}")
    print(f"  Action: {rec['action']}")

Dashboard Access

View analytics in the dashboard:

Navigate to your project in the Dashboard
Click the "Analytics" tab
View Index Health panel with health score and distributions
View Search Analytics with latency, tools, and errors
Use the period selector to view different time ranges

Best Practices

📊

Monitor Weekly

Check your health score and analytics at least once per week to catch issues early.

⚠️

Act on Warnings

Don't let warnings accumulate. Address stale documents and quality issues promptly.

🔄

Enable Auto-Sync

Use GitHub auto-sync to keep documentation fresh automatically with every push.

📈

Track Trends

Compare metrics week-over-week to identify improvements or regressions.

Availability

Index Health & Analytics are available on all plans including FREE. The MCP tools (rlm_index_health, rlm_search_analytics, rlm_index_recommendations, rlm_query_trends) work with VIEWER access level or higher.

View All MCP Tools →