Replay Not Found #1440
Comments
|
Hello! Try install without nginx: |
|
hi @patsevanton thanks for response, I tried to disabled the nginx but I got the same error |
|
This is full values.yaml ? |
|
Try this values: and command for install: |
|
I installed sentry today and I have the exact same issue. The list shows that there should be replays and the data sent by the clients indicated that everything should be working. However I get the error "Replay Not Found" in the UI. What I noticed: This chart is at sentry 24.5.1 - might there be an issue? Is there an easy way to upgrade to a newer Sentry version - this would be interesting in case of security issues as well ... |
|
Tried, still have the same error, here is the image version of different modules (which is by default in helm chart)
posting full values.yaml here |
|
I do not confirm the error install command:
Write the minimum values.yaml with the latest version of the helm chart and the installation command to reproduce the error |
I think there might be a misunderstanding. The error occurs, when trying to view a replay of an issue (or directly via "replays" menu item). When there are no replays, no error occurs.
|
|
Get pod status Try get log by https://github.com/stern/stern: and |
|
@patsevanton here you are, thanks
|
Seems I am not alone : ) |
|
I compare configmap between 23.12.1 and 25.3.0. |
already present in my yaml - using ChartVersion 25.5.0. What I noticed when trying newer image versions with this chart:
|
|
@patsevanton Not work. Deployed the change, and got the same error. |
|
@boindil does the chart work without errors up to version 24.7.1? |
whilst the chartVersion is 25.5.0, the images are version 24.5.1. The default combination does work fine - except for replays. It seems these are sent successfully by the client (HTTP/200 on all calls) but I cannot access these in the sentry UI. I tried with newer image versions up to 24.8.0 with results as described (worse). I created a trial account on sentry.io and only replace the dsn in sentry initialization of my web-app. That works fine. |
|
I am on chart version 25.3.0, and below is the image version of different modules (which is the default in helm chart)
|
|
I am attempting to reproduce an error and have created a repository for this purpose: https://github.com/patsevanton/sentry-react-replay However, I am not receiving any replays. I am not a frontend developer, so I would greatly appreciate it if you could provide a minimal reproducible code example for creating and sending a replay to Sentry. |
Looks good to me at first glance. The DSN has to be replaced in the App.js - it is provided when you create a new project and select in your case react. I‘ll clone your repo tomorrow and give it a shot. |
|
@patsevanton I created a PR on your repo. Only thing you have to do before starting the docker container is to enter your own DSN. Otherwise it will work (show a replay in sentry, but "doesnt find it" with the error we showed). |
|
Additional info: I noticed #1386 having a similar description. Even though clickhouse was the old version that was working there, I upgraded to altinity/clickhouse-server:23.3.19.33.altinitystable (since that one is used in the "original docker variant") via altinity/clickhouse-server:22.8.15.25.altinitystable as seen here: https://github.com/getsentry/self-hosted/blob/master/install/upgrade-clickhouse.sh Not working either :( |
|
Install https://github.com/sentry-kubernetes/charts/releases/tag/sentry-v25.5.0 See replay Press 10.130.0.3 |
|
|
This seems rather strange to me. I would suppose sentry.io runs the same source-code as well, no? Its working there. I did not yet test the docker variant, since my laptop doesnt have enough RAM - will try on my desktop later. The only reference of |
|
the error was reproduced on version 23.5.2 |
|
The error was not reproduced on the latest version in Docker Compose. create virtual machine in internet $ sudo su - get external ip address from command line linux $ curl icanhazip.com $ sudo vim /etc/nginx/sites-available/default $ sudo service nginx restart $ ./install.sh $ cd sentry $ cd .. $ git clone https://github.com/patsevanton/sentry-react-replay.git Build a Docker image: $ docker build -t sentry-example . Launch the container: In the browser, open http://localhost Click on Throw error |
|
Well, thanks for trying with docker, I did not yet get back to my desktop PC .... Could you check out the same sentry version (24.5.1)? getsentry/sentry@0e42a91 This way we would know for sure if it has something to do with the version of sentry or the kubernetes chart here having -some- kind of a problem. |
|
I will check sentry version (24.5.1) on docker-compose |
|
@patsevanton |
|
In progress... The instructions above $ cd self-hosted/ $ ./install.sh i get error on version 24.5.1 and latest commit self-hosted Try |
reference is not a tree: may be ddd9cf25a4e501f00a76186394698f3eeaef2541 |
|
The error was not reproduced on the 24.5.1 version in Docker Compose. create virtual machine in internet $ sudo su - get external ip address from command line linux $ curl icanhazip.com $ sudo vim /etc/nginx/sites-available/default $ sudo service nginx restart $ cd sentry $ cd .. $ git clone https://github.com/patsevanton/sentry-react-replay.git Build a Docker image: $ docker build -t sentry-example . Launch the container: In the browser, open http://localhost Click on Throw error
|
|
@patsevanton do you know which helm chart version works? the current chart does not work with 24.5.1. |
|
welp ... then it definitely seems there is an issue with the kubernetes chart here :/ @patsevanton @Mokto any ideas on what could cause this behaviour? |
I don't know. |
|
@patsevanton |
|
I don't see any extended errors in the extended logs of the new sentry release 25.8.0 |
|
what now? Can we be of any help? |
|
I don't have any ideas. |
|
@patsevanton |
|
@OttoLi-De you're right. But I can't find a reason why help chart doesn't support replay |
|
new helm chart actually made it worse for me - I use sentry in a local cluster with its own CA. the current version doesnt seem to cooperate with that ... getsentry/self-hosted#2950 (happening in web-container as well) |
|
Since which version has self-signed CA support stopped working? |
|
Hi, we have the same issue with the kubernetes helm chart version 25.5.1 on our tanzu k8 cluster. These are our value overrides: And this is the exception which occurs in the senty-web pod: If we find a workarround i will comment it down below. |
|
I encountered the same issue. If you make any progress, please reply to me. Thank you. |
|
|
Can you make a separate issue to discuss the broken self-signed CA support? |
|
So I have digged down a litte bit deeper. I've updated our used helm chart to the latest version |
|
Hi as mentioned here: the Referrer validation errors are "just an annoying log" the main problem lies in the api request for the replay recording segments: I think this error comes from the filestore configuration which is defined in the The configured folder is empty in my pod. Im not yet understanding what should be saved there because the relplays are stored in clickhouse as far is I understand. So why does it even try to read files from there when im loading the segments? I will investigate further... Have a nice day |
|
Good news guys. I finally got it working (sort of). The referrer is not part of the enum errors are indeed just warnings and do not influence the replay storing. It's the For our company storing the replays in s3 is much better than on a pvc so I don't know if i will investigate the root problem with the Have a nice day, |
|
@biclighter81 thank you for your suggestion. I worked out the solution for our case as well. the helm deployment method separates sentry into many services and replay requires few services pointing to the same file storage. To do that, we either need to use s3 or GCS, or when using filesystem, the persistence volume access mode needs to be set as ReadWriteMany I deploy sentry on Azure Kubernetes Service, and to enable ReadWriteMany, I need to attach Azure File as the storage. finally I put the PVC name in existingClaim, and it works. Below are the values I set @patsevanton I think it will be helpful to add to documentation/comment/readme |
|
I'm thinking of thinking about this topic after the open pull request merge and clickhouse update. please tag me next week. |
|
@patsevanton same problem here |
|
@danielehrhardt try |
|
@danielehrhardt did the config example help? |
|
Hey folks, stumbled this by accident :) aside of S3 suggestions mentioned above - in case you're still running into issues please do check if replay request in your browser doesn't die after ~5s Relay's http timeout is 5s by default which might be not enough to load the session back to the client (browser) so you might want to set smtg like this in your helm values: config: # config is top-level key
http:
timeout: 30hope that will help someone 🤞 |
Where to edit this config? |
Issue submitter TODO list
Describe the bug (actual behavior)
I'm not sure if it is a bug or caused by misconfigured. I deployed sentry to the AKS cluster using the sentry helm chart.
I managed to get the issue sent to the sentry instance, however, the replay does not show. I found the following errors in sentry-web.
Expected behavior
Replay can be play in Sentry portal
Your installation details
Steps to reproduce
Trigger error in our website to send an issue with a replay to sentry. However issues in sentry portal shows replay not found.
Screenshots
No response
Logs
No response
Additional context
No response
The text was updated successfully, but these errors were encountered: