From 80eb31cbd8392b9352fc8e034785cc9a14608b04 Mon Sep 17 00:00:00 2001 From: Vlada Chirmicci Date: Fri, 31 Oct 2025 16:40:30 +0000 Subject: [PATCH 1/4] Add footnote about the indexing performance differences between Serverless and ECH Adding more information to explain why single-document indexing in Serverless may appear to be slower than in an ECH environment. Fixes [#385](https://github.com/elastic/docs-content-internal/issues/385). --- .../differences-from-other-elasticsearch-offerings.md | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md b/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md index cd447ace48..fc2798bef5 100644 --- a/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md +++ b/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md @@ -86,7 +86,7 @@ This table compares Elasticsearch capabilities between {{ech}} deployments and S | [**AI Assistant**](/solutions/observability/observability-ai-assistant.md) | ✅ | ✅ | | | **Behavioral analytics** | ❌ (deprecated in 9.0) | ❌ | Not available in Serverless | | [**Clone index API**](https://www.elastic.co/docs/api/doc/elasticsearch/operation/operation-indices-clone) | ✅ | **Planned** | Anticipated in a future release | -| [**Bulk indexing**](/deploy-manage/production-guidance/optimize-performance/indexing-speed.md#_use_bulk_requests) | ✅ | ✅ | The maximum bulk request response time in {{serverless-short}} is 200ms | +| [**Bulk indexing**](/deploy-manage/production-guidance/optimize-performance/indexing-speed.md#_use_bulk_requests) | ✅ | ✅ | The maximum bulk request response time in {{serverless-short}} is 200ms ^1^ | | [**Cross-cluster replication**](/deploy-manage/tools/cross-cluster-replication.md) | ✅ | **Planned** | Anticipated in a future release | | [**Cross-cluster search**](/solutions/search/cross-cluster-search.md) | ✅ | **Planned** | Anticipated in a future release | | **Data lifecycle management** | - [ILM](/manage-data/lifecycle/index-lifecycle-management.md)
- [Data stream lifecycle](/manage-data/lifecycle/data-stream.md) | [Data stream lifecycle](/manage-data/lifecycle/data-stream.md) only | - No data tiers in Serverless
- Optimized for common lifecycle management needs | @@ -103,6 +103,9 @@ This table compares Elasticsearch capabilities between {{ech}} deployments and S | [**Watcher**](/explore-analyze/alerts-cases/watcher.md) | ✅ | ❌ | Use **Kibana Alerts** instead, which provides rich integrations across use cases | | **Web crawler** | ❌ (Managed Elastic Crawler discontinued with Enterprise Search in 9.0) | Self-managed only | Use [**self-managed crawler**](https://github.com/elastic/crawler) | + +^1^ In {{serverless-short}}, data durability is achieved through an object store rather than local replicas. Writes are grouped into batches of 200ms to ensure consistency and persistence, which means that single-document indexing can appear slower than in {{ech}}. However, this design makes {{serverless-short}} more scalable and resilient to high indexing loads without relying on in-cluster replication for fault tolerance. + ### Observability This table compares Observability capabilities between {{ech}} deployments and Observability Complete Serverless projects. For more information on Observability Logs Essentials Serverless projects, refer to [Observability feature tiers](../../../solutions/observability/observability-serverless-feature-tiers.md). From 8960b36ea4e0af63edbfb88da0de93116ca61a84 Mon Sep 17 00:00:00 2001 From: Vlada Chirmicci Date: Tue, 4 Nov 2025 15:40:06 +0000 Subject: [PATCH 2/4] Applying peer review suggestions --- .../differences-from-other-elasticsearch-offerings.md | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md b/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md index fc2798bef5..b28da48c1a 100644 --- a/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md +++ b/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md @@ -86,7 +86,7 @@ This table compares Elasticsearch capabilities between {{ech}} deployments and S | [**AI Assistant**](/solutions/observability/observability-ai-assistant.md) | ✅ | ✅ | | | **Behavioral analytics** | ❌ (deprecated in 9.0) | ❌ | Not available in Serverless | | [**Clone index API**](https://www.elastic.co/docs/api/doc/elasticsearch/operation/operation-indices-clone) | ✅ | **Planned** | Anticipated in a future release | -| [**Bulk indexing**](/deploy-manage/production-guidance/optimize-performance/indexing-speed.md#_use_bulk_requests) | ✅ | ✅ | The maximum bulk request response time in {{serverless-short}} is 200ms ^1^ | +| [**Bulk indexing**](/deploy-manage/production-guidance/optimize-performance/indexing-speed.md#_use_bulk_requests) | ✅ | ✅ | In {{serverless-short}}, Elastic ensures data durability by storing indexed data in an [object store](https://www.elastic.co/blog/elastic-serverless-architecture) rather than local replicas. Writes are batched every 200ms to ensure durability while optimizing performance and cost, which means that single-document indexing can appear slower than in {{ech}}. However, this design makes {{serverless-short}} more scalable and resilient to high indexing loads without relying on in-cluster replication for fault tolerance. | | [**Cross-cluster replication**](/deploy-manage/tools/cross-cluster-replication.md) | ✅ | **Planned** | Anticipated in a future release | | [**Cross-cluster search**](/solutions/search/cross-cluster-search.md) | ✅ | **Planned** | Anticipated in a future release | | **Data lifecycle management** | - [ILM](/manage-data/lifecycle/index-lifecycle-management.md)
- [Data stream lifecycle](/manage-data/lifecycle/data-stream.md) | [Data stream lifecycle](/manage-data/lifecycle/data-stream.md) only | - No data tiers in Serverless
- Optimized for common lifecycle management needs | @@ -103,9 +103,6 @@ This table compares Elasticsearch capabilities between {{ech}} deployments and S | [**Watcher**](/explore-analyze/alerts-cases/watcher.md) | ✅ | ❌ | Use **Kibana Alerts** instead, which provides rich integrations across use cases | | **Web crawler** | ❌ (Managed Elastic Crawler discontinued with Enterprise Search in 9.0) | Self-managed only | Use [**self-managed crawler**](https://github.com/elastic/crawler) | - -^1^ In {{serverless-short}}, data durability is achieved through an object store rather than local replicas. Writes are grouped into batches of 200ms to ensure consistency and persistence, which means that single-document indexing can appear slower than in {{ech}}. However, this design makes {{serverless-short}} more scalable and resilient to high indexing loads without relying on in-cluster replication for fault tolerance. - ### Observability This table compares Observability capabilities between {{ech}} deployments and Observability Complete Serverless projects. For more information on Observability Logs Essentials Serverless projects, refer to [Observability feature tiers](../../../solutions/observability/observability-serverless-feature-tiers.md). From 78c2f6cc2fc0807206b13a31e3e3ef616e57810f Mon Sep 17 00:00:00 2001 From: Vlada Chirmicci Date: Tue, 4 Nov 2025 15:57:53 +0000 Subject: [PATCH 3/4] Add footnote formatting --- .../differences-from-other-elasticsearch-offerings.md | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md b/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md index b28da48c1a..7f70f9673b 100644 --- a/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md +++ b/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md @@ -86,7 +86,7 @@ This table compares Elasticsearch capabilities between {{ech}} deployments and S | [**AI Assistant**](/solutions/observability/observability-ai-assistant.md) | ✅ | ✅ | | | **Behavioral analytics** | ❌ (deprecated in 9.0) | ❌ | Not available in Serverless | | [**Clone index API**](https://www.elastic.co/docs/api/doc/elasticsearch/operation/operation-indices-clone) | ✅ | **Planned** | Anticipated in a future release | -| [**Bulk indexing**](/deploy-manage/production-guidance/optimize-performance/indexing-speed.md#_use_bulk_requests) | ✅ | ✅ | In {{serverless-short}}, Elastic ensures data durability by storing indexed data in an [object store](https://www.elastic.co/blog/elastic-serverless-architecture) rather than local replicas. Writes are batched every 200ms to ensure durability while optimizing performance and cost, which means that single-document indexing can appear slower than in {{ech}}. However, this design makes {{serverless-short}} more scalable and resilient to high indexing loads without relying on in-cluster replication for fault tolerance. | +| [**Bulk indexing**](/deploy-manage/production-guidance/optimize-performance/indexing-speed.md#_use_bulk_requests) | ✅ | ✅ | The maximum bulk request response time in {{serverless-short}} is 200ms [^1^](#footnote-1) | | [**Cross-cluster replication**](/deploy-manage/tools/cross-cluster-replication.md) | ✅ | **Planned** | Anticipated in a future release | | [**Cross-cluster search**](/solutions/search/cross-cluster-search.md) | ✅ | **Planned** | Anticipated in a future release | | **Data lifecycle management** | - [ILM](/manage-data/lifecycle/index-lifecycle-management.md)
- [Data stream lifecycle](/manage-data/lifecycle/data-stream.md) | [Data stream lifecycle](/manage-data/lifecycle/data-stream.md) only | - No data tiers in Serverless
- Optimized for common lifecycle management needs | @@ -103,6 +103,8 @@ This table compares Elasticsearch capabilities between {{ech}} deployments and S | [**Watcher**](/explore-analyze/alerts-cases/watcher.md) | ✅ | ❌ | Use **Kibana Alerts** instead, which provides rich integrations across use cases | | **Web crawler** | ❌ (Managed Elastic Crawler discontinued with Enterprise Search in 9.0) | Self-managed only | Use [**self-managed crawler**](https://github.com/elastic/crawler) | +^1^ $$$footnote-1$$$ In {{serverless-short}}, Elastic ensures data durability by storing indexed data in an [object store](https://www.elastic.co/blog/elastic-serverless-architecture) rather than local replicas. Writes are batched every 200ms to ensure durability while optimizing performance and cost, which means that single-document indexing can appear slower than in {{ech}}. However, this design makes {{serverless-short}} more scalable and resilient to high indexing loads without relying on in-cluster replication for fault tolerance. + ### Observability This table compares Observability capabilities between {{ech}} deployments and Observability Complete Serverless projects. For more information on Observability Logs Essentials Serverless projects, refer to [Observability feature tiers](../../../solutions/observability/observability-serverless-feature-tiers.md). From 4cb82eadacad22dba4f3497b8be950480f8ac1c1 Mon Sep 17 00:00:00 2001 From: Vlada Chirmicci Date: Wed, 19 Nov 2025 12:10:01 +0000 Subject: [PATCH 4/4] Implementing review feedback --- .../differences-from-other-elasticsearch-offerings.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md b/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md index 7f70f9673b..7991b693aa 100644 --- a/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md +++ b/deploy-manage/deploy/elastic-cloud/differences-from-other-elasticsearch-offerings.md @@ -86,7 +86,7 @@ This table compares Elasticsearch capabilities between {{ech}} deployments and S | [**AI Assistant**](/solutions/observability/observability-ai-assistant.md) | ✅ | ✅ | | | **Behavioral analytics** | ❌ (deprecated in 9.0) | ❌ | Not available in Serverless | | [**Clone index API**](https://www.elastic.co/docs/api/doc/elasticsearch/operation/operation-indices-clone) | ✅ | **Planned** | Anticipated in a future release | -| [**Bulk indexing**](/deploy-manage/production-guidance/optimize-performance/indexing-speed.md#_use_bulk_requests) | ✅ | ✅ | The maximum bulk request response time in {{serverless-short}} is 200ms [^1^](#footnote-1) | +| [**Bulk indexing**](/deploy-manage/production-guidance/optimize-performance/indexing-speed.md#_use_bulk_requests) | ✅ | ✅ | The baseline write latency in {{serverless-short}} is 200ms [^1^](#footnote-1) | | [**Cross-cluster replication**](/deploy-manage/tools/cross-cluster-replication.md) | ✅ | **Planned** | Anticipated in a future release | | [**Cross-cluster search**](/solutions/search/cross-cluster-search.md) | ✅ | **Planned** | Anticipated in a future release | | **Data lifecycle management** | - [ILM](/manage-data/lifecycle/index-lifecycle-management.md)
- [Data stream lifecycle](/manage-data/lifecycle/data-stream.md) | [Data stream lifecycle](/manage-data/lifecycle/data-stream.md) only | - No data tiers in Serverless
- Optimized for common lifecycle management needs | @@ -103,7 +103,7 @@ This table compares Elasticsearch capabilities between {{ech}} deployments and S | [**Watcher**](/explore-analyze/alerts-cases/watcher.md) | ✅ | ❌ | Use **Kibana Alerts** instead, which provides rich integrations across use cases | | **Web crawler** | ❌ (Managed Elastic Crawler discontinued with Enterprise Search in 9.0) | Self-managed only | Use [**self-managed crawler**](https://github.com/elastic/crawler) | -^1^ $$$footnote-1$$$ In {{serverless-short}}, Elastic ensures data durability by storing indexed data in an [object store](https://www.elastic.co/blog/elastic-serverless-architecture) rather than local replicas. Writes are batched every 200ms to ensure durability while optimizing performance and cost, which means that single-document indexing can appear slower than in {{ech}}. However, this design makes {{serverless-short}} more scalable and resilient to high indexing loads without relying on in-cluster replication for fault tolerance. +^1^ $$$footnote-1$$$ In {{serverless-short}}, Elastic ensures data durability by storing indexed data in an [object store](https://www.elastic.co/blog/elastic-serverless-architecture) rather than local replicas. Writes are batched over a 200ms window to ensure durability while optimizing performance and cost, which means that single-document indexing can appear slower than in {{ech}}. However, this design makes {{serverless-short}} more scalable and resilient to high indexing loads without relying on in-cluster replication for fault tolerance. Because of a higher baseline write latency, {{serverless-short}} indexing can be scaled by increasing concurrent indexing clients. ### Observability