Hi Daniel, to assist our investigation can you please tell if multiple users see this issue, or it’s only your user seeing this error? Can you please also clarify the flow - Does this issue happen daily after uploading new data to your workspace? Or this happens sporadically and not only first thing in the morning when you login? And could you DM a direct link to the dashboard please so our team can review it, if needed. Thanks in advance for this!

I would just like to add that I went ahead and accessed your workspace and used Impersonation on our internal Admin user and I was able to replicate the issue on my side

Hello again Daniel, another follow up: I've discussed your issue internally, it comes out that we might be dealing with some bug related to JDBC driver for Amazon RedShift. Error in the filter display is related to failed execution of SQL statement. We were able to replicate it and currently we are working on a fix, I will inform you about any development of this case. Could you please tell me what is the business impact caused by this issue?

and one more question, I've noticed that query execution fails when below query is executed:

SELECT "app_sys_apps_control_list0"."app" AS "a_label_apk_app_selection_app_system_2bb0c93f87e411", "t1"."m_f567ed6a0961b25176acfb49f8490b11" AS "m_1"
FROM (SELECT "apk" AS "a_label_apk_app_system_12bd071c48e760", COUNT("apk") AS "m_f567ed6a0961b25176acfb49f8490b11", TRUE AS "def_m_f567ed6a0961b25176acfb49f8490b11"
FROM "local_schema_rs_gd_v10_cluster3"."app_sys_apps_control_list"
GROUP BY "apk") AS "t1"
INNER JOIN "local_schema_rs_gd_v10_cluster3"."app_sys_apps_control_list" AS "app_sys_apps_control_list0" ON "t1"."a_label_apk_app_system_12bd071c48e760" = "app_sys_apps_control_list0"."apk"
WHERE "t1"."def_m_f567ed6a0961b25176acfb49f8490b11"
ORDER BY "app_sys_apps_control_list0"."app" NULLS FIRST
LIMIT 10000

Could you please test that query against your data in your environment?

hi, confirming that this query actually works within 2 seconds when applied directly to my db

Hello Daniel, Coming back to our investigation. We would have some additional questions: 1. Following instructions from https://repost.aws/knowledge-center/redshift-query-abort, could you please verify ABORT, CANCEL, or TERMINATE requests for date frame from 26th of July until today. 2. Also, due that logs are stating “ProcessError(errorType=SERVER_EXCEPTION, message=No connection to '5e980bdc-1165-422f-8036-4bee8a5106ea' available.)” - we see that connection to the server was interrupted several times (we can’t identify on which side it was interrupted), it happens intermittently - it could be due to firewall configuration which stops some particular connections when discovered a malicious content of network packets (that is how stateful firewalls are behaving). It might also related to poor connection between GoodData Cloud and your RedShift database. 3. When the issue occurred first time? Where there any changes done your side at that time? 4. What is the quota for maximum concurrent connections to the database and their timeout set on RedShift? Last but not least I would like to specify urgency for this incident with you. Please let me know how impactful is this problem for your business. Kind regards, Jakub

Thank you for your reply. We have increased the urgency of your incident and engaged additional resources in investigation. Please accept our apologies but due to complexity of the issue. Just to give you a short overview of our investigation: • as previously, also today I've impersonated to your affected workspace - however I was unlucky to reproduce the problem probably due that data for filtered value was already precached • we investigated jdbc driver logs in context of current driver version to verify if there isn't any incompatibility • we created a test workspace with RedShift datasource and dummy database to replicate the error - we did not get same error. • we analyzed "SELECT" query which is executed when there is input in filter field and compared with ones we executed while impersonating and also in our separate test environment We will much appreciate if you could provide a new HAR file capturing inputs in the filter field, if I will be able to replicate the issue later on today with second impersonation attempt then I will let you know that HAR file is not necessary anymore. Also I'm preparing additional investigation on your serverless RedShift to check in Query Editor reason of SQL query failure. Best regards, Jakub

Hi Daniel, we are in active discussion with AWS, we have provided a reproducer and they escalated to their Redshift Services team. I will keep you posted.

Hello @Daniel Muise, AWS has concluded that the issue is a limitation of AWS Redshift Serverless, not a bug. We are planning to implement a retry mechanism on our side. While this may slightly prolong the initial computation of reports when Redshift Serverless is waking from sleep, it will prevent failures. This enhancement is expected to be implemented within the upcoming month.