[
https://issues.apache.org/jira/browse/HDDS-17593?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
rajesh updated HDDS-17593:
--------------------------
Description:
An SCM trace today is the RPC span. A key create shows AllocateScmBlock, and a
key read shows GetContainerWithPipelineBatch, with no breakdown of block
allocation, container selection or SCM Ratis, and no failure status on the span.
Close that gap:
* Block allocation, the client and block protocol servers, the block manager,
the Ratis server and writable container selection show up under the RPC span.
* A failed SCM request records the request type, the status and the exception.
* allocateBlock records the requested size and count, the chosen container and
pipeline, and the failure when allocation cannot find a container.
Depends on HDDS-17581. Reuses the same tracing helpers as HDDS-17586. No
separate proxy, attribute or local-span sub-tasks.
was:
Improve OpenTelemetry tracing in SCM so that a trace shows the work behind
block allocation, container and pipeline operations, and SCM Ratis, not only
the RPC span. The work reuses the \{{hdds-common}} tracing building blocks
added for OM.
The work is split into three sub-tasks:
# Create tracing proxies for SCM components.
# Add span attributes for SCM requests.
# Add logical spans for critical paths that the proxies don't cover.
> Tracing improvements for Storage Container Manager
> --------------------------------------------------
>
> Key: HDDS-17593
> URL: https://issues.apache.org/jira/browse/HDDS-17593
> Project: Apache Ozone
> Issue Type: Task
> Reporter: rajesh
> Assignee: rajesh
> Priority: Major
>
> An SCM trace today is the RPC span. A key create shows AllocateScmBlock, and
> a key read shows GetContainerWithPipelineBatch, with no breakdown of block
> allocation, container selection or SCM Ratis, and no failure status on the
> span.
> Close that gap:
> * Block allocation, the client and block protocol servers, the block manager,
> the Ratis server and writable container selection show up under the RPC span.
> * A failed SCM request records the request type, the status and the exception.
> * allocateBlock records the requested size and count, the chosen container
> and pipeline, and the failure when allocation cannot find a container.
> Depends on HDDS-17581. Reuses the same tracing helpers as HDDS-17586. No
> separate proxy, attribute or local-span sub-tasks.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
