Skip to content

Handle errors in a sane way #14

@davidwin93

Description

@davidwin93

Our current error handling should be improved to handle failures more gracefully. We currently retry failed attempts endlessly and we also endlessly retry Gitlab operations (update comment, etc...) . Adding some backoff with a max attempts would be a good start. We could also augment our Queue messages to maintain a count of attempts to prevent endless retries.

For NATS we could use some combo of these to manage message delivery:
https://pkg.go.dev/github.com/nats-io/nats.go#MaxDeliver
https://pkg.go.dev/github.com/nats-io/nats.go#Msg.Term
https://pkg.go.dev/github.com/nats-io/nats.go#Msg.Metadata

@sl1pm4t pointed to this https://choria.io/blog/post/2020/04/03/nats_patterns_9/ under Redelivery as a way of handling these types of messages.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions