Skip to content

Recording BackendTime incase of failure scenario #663

@PrathibaJee

Description

@PrathibaJee

As per the requirement , the backend-time is the
"Value written by the service provider, reporting the elapsed time for data retrieval from the backend (service invocation, database access…). Expressed in milliseconds"

In version 1.0.0 of ApplicationPattern , there was no support for a database.
Whereas from version 2.0.0 , we have ElasticSearch as a backend for the applications EATL , OL and ALT.
So , the time spent in the backend is now included in the header parameter backendTime.

Please find the details of the scenarios for which we are recording and not recording the BackendTime ,

Scenarios isRecorded? comment
A successful Single call to ES yes
A successful Single call to ES , but an internal error due to server side logic . no
A REST service with multiple calls to ES. And all calls are successful yes consolidated time of each call to ES
A REST service with multiple calls to ES. And if any one call to ES is failed no
A REST service with multiple calls to ES. And all calls are successful. but an internal error due to server side logic . no

In a highlevel , in the server side implementation , for negative scenarios , we are not recording the backendTime and it will be 0.
@DanaSunal , kindly please provide your input aswell.

Kindly please let us know your views.
If this implementation is fine, then we also need to document it in the specification. Thanks.

Reference : openBackhaul/ApplicationLayerTopology#298 (comment)

Metadata

Metadata

Assignees

Projects

No projects

Relationships

None yet

Development

No branches or pull requests

Issue actions