Closed thoven87 closed 4 weeks ago
Here are some initial comments. In general this look good, I was wondering if it is worthwhile adding a metric for the retry count.
For another PR if we added a dateCreated to the QueuedJob, we could add a time job queued metric. In a way dateCreated is probably a useful piece data regardless.
Thanks, for the feedback. I added a retry counter. I will open another PR to add both createdAt
and maybe Job name? Right now, Job.id.name
does not always exist in every context. In some context, it's just Job.id which is a UUID.
Looks great! Just one final bit of consistency
I think this Pull request is ready. Below is a dashboard I created to test this implementation.
Next step would be to add tracing and queued at to compute avg wait time.
also previous implementation of the dashboard to show the failed metrics
This is great
Please merge at your convenience! What's the best place to share RED metrics dashboards? I am in the process of creating a few Grafana dashboards for Hummingbird.
There would be three dashboards in total:
1 - Display the overall health of an application including jobs and http metrics, spans and error logs. 2 - Display Http metrics, spans and error logs 3 - Display Jobs metrics, spans and error logs
Just waiting on a final approval from @Joannis
@thoven87 I'd love to post them either in a docs section, on docs.hummingbird.codes as a DocC guide. I think we can also highlight it on swiftonserver.com
@thoven87 I'd love to post them either in a docs section, on docs.hummingbird.codes as a DocC guide. I think we can also highlight it on swiftonserver.com
@Joannis I suppose for DooC guide it would need to be added in a section in this repository Indeed, it would be great to highlight it on swiftonserver.com as well.
You could possibly include something in the observability docs but https://docs.hummingbird.codes/2.0/documentation/hummingbird/loggingmetricsandtracing.
I think an end to end metrics and tracing article in swiftonserver.com which uses these would be a great addition.
You could possibly include something in the observability docs but https://docs.hummingbird.codes/2.0/documentation/hummingbird/loggingmetricsandtracing.
I think an end to end metrics and tracing article in swiftonserver.com which uses these would be a great addition.
Thanks Adam, I will take a look and see what I can do.
Here's my first attempt at adding Observability to the Swift jobs project.
I would love some feedback on how to make metrics better. cc @adam-fowler
Since everyone might use different tools to display their dashboards, I will attempt to provide a grafana dashboard as it is what I am most familiar with.
Lastly, most of the tests are a compilation of the current job tests and Hummingbird metrics tests.