Azure AI Search Indexer
Does ZopNight manage Azure AI Search Indexer?
Search indexers pull data from sources into indexes on a schedule, and every run through a skillset spends money on AI enrichment. ZopNight records each indexer's schedule and status via the service admin key and flags high-frequency indexers. An indexer firing every 5 minutes re-bills enrichment 288 times a day.
Rules that fire on Azure AI Search Indexer
No active rule family targets Azure AI Search Indexer today. Rules that used to are retired, and retired rules publish no pages and fire no findings. Scheduling and permissions coverage are unaffected.
At a glance
| Field | Value |
|---|---|
| Scheduling notes | discovery only. |
Indexers automatically pull data from sources into search indexes on a schedule. Aggressive indexer schedules drive skillset (AI enrichment) charges on every run.
Indexer schedules are a cost multiplier
An indexer is free to define and free to leave idle. Its cost behavior comes from what each run triggers: documents flow from the data source, and when a skillset sits in the path, every processed document incurs per-transaction AI enrichment charges on the attached AI Services account. The schedule multiplies that. A crawl interval chosen casually during development, every few minutes to make the demo feel live, becomes hundreds of enrichment-billing runs per day in perpetuity, on data that may change once a week.
Schedule and status via the admin key
Discovered via the search enricher with schedule and status detail. Indexers are data-plane objects on the service endpoint, not ARM resources, so ZopNight enumerates them with the service’s primary admin key alongside the indexes, skillsets, and data sources they connect. The captured schedule is what lets high-frequency indexers be flagged, and the status field separates healthy pipelines from ones failing silently. Indexers are inventory rows only, and the fix is editing the schedule or disabling the indexer in the service, not a platform stop verb.
Runaway indexer patterns
The expensive findings repeat across estates: a demo-era five-minute schedule feeding a skillset, quietly the largest AI Services line in the resource group; indexers left enabled against sources that stopped changing months ago, re-scanning on every tick; and failed indexers in a retry loop, paying enrichment for partial batches while never completing. A fourth is duplication: parallel indexers built by different teams over the same source, enriching the same documents twice into sibling indexes.
Indexer runs in the portal
Azure portal → AI Search → select the service → Search management → Indexers shows each indexer with its schedule, last run, and success or failure state. The execution history per indexer is the quickest way to see how often a pipeline actually fires, and what each firing has been costing in enrichment.