It's here @oliver.drewes_gitlab https://gitlab.com/glitchtip/glitchtip-backend/-/blob/master/glitchtip/settings.py
I use whitenoise for static files and django-storages for media.
I'm not sure exactly what you're looking to achieve. I know GT works fine with volumes in docker compose and fine with external media on k8s.
Got it, it should be possible with some work on settings. I can't think of any fundamental reason it wouldn't be. I'm not sure if you'd have to modify the frontend or not as well. Happy to review any merge requests.
I could see the frontend being an issue to solve. The base url would change and it's not acceptable to set that during build time.
Do you think a rewrite at boot would help by changing the base path? Like with a sed of the index file?
Found this article as the only appropriate result on my search: https://pretagteam.com/question/how-to-set-application-context-path-in-angular2-routing-properly
We are using log4j2 based sentry appender for glitchip everything is working seamlessly. One of the issues we see is that the grouping of the errors is not happening per template message, instead of that grouping is happening on the exact log message. Has anyone seen this and any way to resolve the same?
Example:
With log4j2, it’s expected that each log template will be grouped together. But not each time different variables values are used.
logger.info("User {} has logged in using id {}", map.get("Name"), user.getId());
In this case there should be 1 issue, regardless of the value of name and id.
Regarding base path - sed sure seems not ideal. A nice thing about our current set up is that the angular bundle is just a static bundle and doesn't know anything about Django. And Django doesn't need to know anything about angular. It just serves it. That very nice simplicity is not great if you want to pass information such as base path from a Python env var to Angular. I'm open to proposals. I understand the use case of having a variable base path.
Regarding how we group events @chethan - I believe the message string is generated by older sentry open source code ported to python 3. If my memory is correct, the grouping of events is very simple and based on message. You could look at how the old open source sentry handles your events for comparison. If you like the way OSS Sentry does it more, you could port that code or find out why it doesn't work already. Though we don't strictly have to follow sentry here as it wouldn't have anything to do with API compatibility. I would probably lean towards a good default that could apply some level of fuzzy logic on matching like events. But I'd defer to anyone willing to put in the work. I think Sentry had some more options on ways this logic could be customized but I never really used that.
I am pleased to announce that GlitchTip 1.9, introducing frontend support for Uptime Monitoring, has been released:
https://glitchtip.com/blog/2021-11-30-glitchtip-1-9
https://glitchtip.com/blog/2021-11-30-glitchtip-1-9
How do people use notifications in the real world? Any feature requests? Uptime monitoring has very limited notifications in this initial feature release.
We're eventually making some backwards incompatible changes to how we run celery in docker. The goal is to use read only docker containers. If you are use docker compose, please ensure your worker command is ./bin/run-celery-with-beat.sh instead of celery -A glitchtip worker -B -l INFO
We'll post this on the blog later too.
If anyone ever wants to contribute translations, I would add code to support it. Don't think the coding would be tooo hard. At least to get started.
Greetings, @jegade (Jens Gassmann). If you are running your own instance, you can search by user email in the admin backend.
How could i format the search string?
ah, ok, email/id direkt into the search field
the first time this does not work :-(
Are you talking about a issue reporting end user? As in the set user who experienced an error. I think James might have been referring to GlitchTip internal users.
For any search issue, we index most event data (which itself is stored in a jsonb field). So if it doesn't come up it's either
- Didn't ingest the data as expected
- Something "weird" about how the data is indexed
- Something off about parsing the search query string
- A rare case where the data isn't in the jsonb field
- Celery index task isn't running
So you could debug it in that rough order. Make sure the data is in the database at all.
Hope that helps. If you think you found a bug or have an idea for improvement you should open an issue and possibly merge request
Hi. I'm running self-hosted GlitchTip on Kubernetes with AWS RDS for Postgres DB.
All is running fine except that I am noticing a rapid increase in the size of the search_vector index of the issues_issue table without any increase in ingestion. My retention period is also very small at 10 days. The size of the index was fine for the first couple of weeks but since then it has started growing at a rapid pace and no matter how much the size of DB is increased, it consumed all the storage. The size of the issues_issue table is very small though but the index is growing crazily.
Any insights on this behaviour?
All is running fine except that I am noticing a rapid increase in the size of the search_vector index of the issues_issue table without any increase in ingestion. My retention period is also very small at 10 days. The size of the index was fine for the first couple of weeks but since then it has started growing at a rapid pace and no matter how much the size of DB is increased, it consumed all the storage. The size of the issues_issue table is very small though but the index is growing crazily.
Any insights on this behaviour?
Ideas for easier contributions
- Arm builds - docker knowledge needed - glitchtip/glitchtip-backend#88
- Pretty basic work around Django Rest Framework glitchtip/glitchtip-backend#107
- Update documentation, requires only editing markdown https://gitlab.com/glitchtip/glitchtip-frontend/-/tree/master/src/sdk-docs pick your favorite language/framework and make sure the instructions are up to date and helpful. Django docs are a better maintained one to reference
i can see that can use sentry and thats great but if im starting out on a new app i dont have to
Hi Deepak! I'm developing as we speak, so yes we are definitely active.
and i got the code in sentry repo in glitchtip org
rate_limit_count
i don't mind contributing but i should know wether its active or not
shall we use it in our org or move to something else
shall we use it in our org or move to something else
plus i see no replys to recent issues
hence i asked its still active or not
I've been focusing on uptime monitoring recently, but we do have plans to devote some time to outstanding issues.
And, of course, everyone is welcome to contribute on issues they find more pressing.
as i didn't saw the code for it
I haven't worked with rate limits myself. It should be pretty easy to test with one of our error factories, though.
If you run into any problems with it, though, let us know.