You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Since we have 2 replicas, fetching data for inbound connectors become inconsistent. Sometimes we fetch data from one replica, sometimes from the other, causing strange changes in the Webhook/Subscription tab in web-modeler.
While investigating I identified an other issue/improvement opportunity: we do not distinguish between logs coming from cluster1 and cluster2. We always merge all the incoming logs and show them together. It would be nice to show the logs separated by clusters. Separation by nodes on the other hand might not make much sense if we somehow identify the log entries (e.g. based on timestamp) and only add new ones but never delete any. This would ofc also require a guarantee that we always have info from all nodes. Regardless, I've reached out to console and sre team about this to have the whole picture.
Describe the Bug
Since we have 2 replicas, fetching data for inbound connectors become inconsistent. Sometimes we fetch data from one replica, sometimes from the other, causing strange changes in the Webhook/Subscription tab in web-modeler.
Live example: https://modeler.ultrawombat.com/diagrams/b972d768-7208-4b48-a9e1-69bc819ba99a--kafka-incorrect-activation-condition?v=760,386,1
Modeler._.Kafka.incorrect.activation.condition.-.Google.Chrome.2024-08-02.11-37-58.mp4
Steps to Reproduce
Expected Behavior
We should not have the duplicated data
The text was updated successfully, but these errors were encountered: