Closed j82w closed 1 month ago
and can help show thread starvation for requests ran in parallel like strong reads
How can this be used? Is the Start Time captured before the Task is created or after? If it's before, should this then be compared with the Create on the request timeline?
How can this be used? Is the Start Time captured before the Task is created or after? If it's before, should this then be compared with the Create on the request timeline?
The start time gets set before the task is created. I updated the description.
Pull Request Template
Description
This PR adds the store result duration time to the diagnostics. The information was already being captured but was not added to the JSON diagnostics. This is useful to help determine how long each individual request to the backend took and the duration between each call.
Example of a strong write. This shows the initial write is response time is 2024-05-01T16:01:48.4597873Z. The barrier request response time is 2024-05-01T16:01:48.6634187Z. This gives 204ms between response time and the barrier request took 86ms leaving 172ms between the write response and the start of the barrier request. Before the diagnostics only had the transport start time so it was not clear where the latency was coming from.
Type of change
Please delete options that are not relevant.
Closing issues
To automatically close an issue: closes #IssueNumber