Managing Context Grounding indexes

Context Grounding index monitoring on the Agents page, covering query and ingestion metrics, per-index details, and the automated index health score.

The Context grounding indexes tab enables you to monitor the status and health of your indexes and track their ingestion progress at the tenant level.

The Index Monitoring section provides graphical insights into the following metrics:

• Queries: Tracks the number of queries performed on your indexes over a selected time interval.
• Index jobs: Displays the number of completed ingestion jobs. If no jobs have run in the selected period, a placeholder graphic is shown.

The Indexes table lists all available indexes in your tenant. For more granular information on indexes, select an index from the table. The individual view provides the following information:

• Index name, Folder, and Data source
• Health status, indicating whether an index is Healthy or Degraded
• Storage size, Ingestion status, Last sync, and Last queried timestamps

Select an index from the table to open its details page. The Index details view provides:

• Queries during the selected interval.
• A Health chart showing the overall health score and the two component criteria.
• Ingestion history, listing completed, failed, skipped, or deleted sync runs.

Use this view to investigate degraded indexes, review ingestion reliability, or check when an index was last queried.

Index health score

The Index health helps you monitor the condition of your Context Grounding indexes and identify those that may need maintenance or cleanup. It provides an automated, objective score showing how well each index performs and how frequently it’s used.

Each index receives a health score based on two criteria:

• Ingestion reliability – How successfully documents were processed during the latest ingestion run.
• Utilization – How recently the index has been queried.

The combined score helps you determine whether an index is healthy and delivering value (overall score above 75%), or degraded (overall score 75% or below). The score updates automatically after each successful ingestion or query.

Index health is recalculated every 15 minutes for all tenants and only applies to indexes with completed ingestion jobs. New indexes have a 24-hour grace period after creation before health scoring begins.

The following table describes how each health criterion is calculated and weighted to determine the overall index health score.

Criterion	Metric	Formula	Notes
Ingestion reliability	Success rate from last ingestion	100 × (1 - failed_docs / total_docs)	Skipped and deleted documents are excluded. If no ingestion is completed within 24h, score = 0.
Utilization	Days since last query (max 90 days)	100 × (1 - days_since_last_query / 90)	After 24h without any query, score = 0.
Overall health	Average of both metrics	(Ingestion Reliability + Utilization)	Calculated only if both component scores exist.

You can review the health status in two places:

• The Context Grounding Indexes list – The Health column shows a colored indicator (green for Healthy, red for Degraded). Hovering over it displays details for each dimension (Ingestion reliability and Utilization).
• The Index details view – The dedicated Health chart displays the overall score and component metrics. You can also track ingestion history and investigate issues when reliability is low.