Skip to content

Optimize performance of updating ingest pipelines when operation is a no-op #77382

Closed
@joshdover

Description

@joshdover

When Fleet installs and upgrades packages, we currently run a PUT on every asset we need to write to Elasticsearch. Often, package assets will not change from one version to the next.

Elasticsearch has already optimized no-op updates to index and component templates by avoiding a cluster state write when the template has not changes. This results in single-digit millisecond response times on these operations.

Ingest pipelines do not yet have a similar optimization and this causes some delay that may be avoidable.

Index Template on no-op PUT

image

Ingest Pipeline on no-op PUT

image

Related to elastic/kibana#110500

Metadata

Metadata

Assignees

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions