is kent useful to you?

[willkg]

I get zero signal for the work I do and have no idea if it's generally positive or negative in the world. I do a lot of work, so this is weird.

I'm experimenting with ways to get some signal with the least amount of effort for you.

If you have some positive feeling about Kent (it's useful! you think it's a cool idea! you with you named your project Kent!), add a 😄 or a +1 or something with the thing in the upper right hand corner.

If you have some negative feeling about Kent, add a :frown: or a -1.

[timfish]

I created something similar in JavaScript for testing the Sentry Electron SDK e2e tests:
https://github.com/getsentry/sentry-electron/blob/master/test/e2e/server/index.ts

It works but my main concern has been that this is not actually testing against real Sentry endpoints.

To improve on this I've been trying to get Sentry's Relay server working as a test server. Fortunately, it already had mode that did something similar so the change required were minimal. I've got this PR which exposes an http endpoint for querying the latest events:
getsentry/relay#1151

I then created a minimal node.js library that wrap the server process and polls for events:
getsentry/sentry-javascript#4254

If I can get the Relay changes merged, you could wrap the server process in Python too!

[willkg]

I don't necessarily need something written in Python. If it's a service, then I can stick it in a Docker container and it doesn't really matter what it's written in as long as it exposes an API that I can query and assert statements against.

I looked at Relay and thought about having Kent sit on the receiving end of Relay. So then you could have a Docker container running Relay and Kent and your application would set the SENTRY_DSN to point to Relay and Relay would point to Kent.

I didn't end up going with that because I don't need to test and verify the sentry-sdk is working. I only needed to test whether the sentry-sdk is set up in my application correctly to capture errors and send them to a sentry instance and whether the data in the error events is being sanitized correctly.

I'll keep tabs on your changes, though. It's definitely interesting.

[diox]

FWIW I just successfully used kent to debug before_send configuration locally without having to set up an actual Sentry service locally. I liked the ease of install and low footprint, very easy to set up/tweak just for some quick hacking.

Real testing will be against an actual Sentry server, but this was perfect for development. In the past I had set up a full onpremises Sentry instance and while that worked, it was a lot more time consuming.

[yoobles]

This project is very useful! We use Sentry in production and having a lightweight fake Sentry server for local development is very useful. It's not for collection of any exceptions but more so that our code can be written with more of a guarantee that some form of Sentry will exist.

[dimaqq]

I'm going to try it :)

[dimaqq]

I've pushed it to my own docker repo as quay.io/dimaqq/kent

[kobus-v-schoor]

This is great project, thanks for maintaining it! Was super simple to get up and running and makes debugging sentry stuff a breeze. Might I suggest publishing a docker image to docker hub via Github actions on each release (I'd be happy to open a PR with the required workflows), would make it even easier to run as part of a docker-compose stack.

EDIT: PR made here: #47

[rexzhang]

I made a docker image release for myself, hopefully it will be useful for others as well!