Ianmacleod/error handling #282
Conversation
…ngine into ianmacleod/error_handling
… unneccessary logging errors
There was a problem hiding this comment.
adding this configuration because we are currently seeing error logs related to this not being properly set up, the error logs go away when this is off. ref: DataDog/dd-trace-py#5212
yixu34
left a comment
yixu34
left a comment
There was a problem hiding this comment.
Just wanted to clarify the repository field
| detail="deletion of endpoint failed, compute resources still exist.", | ||
| ) from exc | ||
| except Exception as exc: | ||
| request_id = get_request_id() |
There was a problem hiding this comment.
Think this is fine for now, but as we revisit self-hosting in VPCs, we'll probably want some sort of Gateway abstraction for getting the request/trace ID.
| Thrown when a user tries to specify a custom Docker image that cannot be found. | ||
| """ | ||
|
|
||
| repository: str |
There was a problem hiding this comment.
Just want to check that repository refers to a full image URL, minus the image tag. e.g. it shouldn't refer to an AWS ECR repo purely by its name. In other words, it shouldn't look like some_service, but <account_id>.../some_service or <some dockerhub prefix>.../some_service.
There was a problem hiding this comment.
Ok discussed offline, this is fine for now. If needed in the future, we can deprecate this field and come up with something new in its stead, like image_url.
Creates structured logging instead of having stack traces spread out over multiple logs.
Consolidates exceptions into one unified place.
Adds support for
request_idthat is surfaced when an uncaught exception occurs.