ElasticSearch Performance metrics

[mn_24]

Hi,

We'd like to achieve the following: to have metrics for alerting based on:

### Search Performance (`GET /_nodes/stats/indices/search`)
| `query_time_in_millis / query_total` (avg latency) | > 100ms | > 500ms |
| `fetch_time_in_millis / fetch_total` | > 50ms | > 200ms |

### Indexing Performance (`GET /_nodes/stats/indices/indexing`)
| `index_time_in_millis / index_total` (avg) | > 50ms | > 200ms |
| Indexing rate drop vs baseline | > 50% drop | stalls for > 1 min |

What we've done so far is not exactly what we are expecting and just to mention that we are in Managed and not in SaaS!

[Mohamed_Hamdy]

Hi,

Assuming you're using the Elasticsearch Extension, I would suggest checking the node-es-operations feature set, which provides the required Query, Fetch, and Indexing time/count metrics, along with Indexing Rate.

You could use Data Explorer in Advanced mode to calculate average latencies and then configure Metric Events for alerting.

Just ensure the calculations account for cumulative counters and use the correct monitoring intervals.

For indexing rate drops, you could explore adaptive or seasonal baseline alerting, although an exact 50% drop may require additional logic.

Could you share what you've configured so far and what's not working as expected?