Sense System Performance Analyzer connection error

[covenant_bi]

Hello, we are trying to set up the Sense System Performance Analyzer but we are getting the same error whether we run the reload in the Hub or in QMC.  

Some particulars about our deployment:

  central and 1 scheduling RIM node, all other monitor apps reload successfully

  Qlogs and other  REST connectors use FQDN of central server in connection string

  no other virtual proxy set up in the environment

The following error occurred:
Connector reply error: Unable to connect to database:
Error message:
Please check the values for Username‚ Password‚ Host and other properties.
Description: Communication link failure
The error occurred here:
SELECT
"id",
"entry_timestamp",
"process_host",
"entry_level",
"event_message",
"event_id",
"event_source",
"event_host_name"
FROM "public"."view_platform_events"
WHERE entry_timestamp >= '2020-05-08 06:44:52.000'

[SaudZ]

I get the exact same error. All other monitoring apps are working fine (overwriting the missing database credentials) but this monitoring app, Sense System Performance Analyzer, does not finishing running. 

Anyone found a solution yet?

[covenant_bi · Accepted solution]

Hello, I did end up opening a ticket with Qlik.  They had me try a bunch of fixes that didn't work.   Ultimately, they wanted me to delete all my monitoring apps and re  config them all.   I didn't want to do that.  So I started digging and found that the error was referring to database communication. So I started looking into Postgres issues and communication.  There are some files that you work with on setup. Referenced here: https://community.qlik.com/t5/Qlik-Support-Knowledge-Base/PostgreSQL-postgresql-conf-and-pg-hba-conf...

I found that our pg_hba.conf file didn't have the right settings as far as whitelisted IP addresses.  The file is located here:

C:\ProgramData\Qlik\Sense\Repository\PostgreSQL\9.6

According to the link above, I tried using 0.0.0.0 and the standard Qlogs connection built in.  I still didn't see any success.  So I explicitly entered the IP of my central node (we have a scheduling node too) and tried standard Qlogs_reader user connection again..still nothing.  Finally I decided to create a new Postgres connection (same as Qlogs) with the same Qlogs_reader user and password as the other Qlik released connection. I was able to test connect successfully, so I replaced the Qlogs LIB connects in the Sense System Performance Analyzer app and it completed successfully!

Also, I was noticing the same error in the License and Monitoring app (when ran in Hub mode) but those apps ignore the error and continue running.  When I replaced the Qlogs LIB connects in those apps, they ran with no errors at all. 

Qlik couldn't explain why, but I now have the whole suite of Qlik released monitoring apps running successfully.  Please let me know if you have any questions. 

[SaudZ]

Thank you for the swift reply with thorough instructions.

We have removed centralised logging and logging database (QLog). I was wondering, have you done the same? 

[covenant_bi]

No we have it set up, should we be running off the log files vs db logging?  Better efficiency?

[SaudZ]

We noticed it was taking up a lot of space and it looks like you cannot unselect the "physical" log files so we went for the only possible option, removing the logging in the db.

[pauloandre]

Qual senha do usuário Qlogs_reader foi utilizada para poder recriar uma nova conexão??

E o que deve ser feito de diferente nessa nova conexão para funcionar??

Obrigado

[covenant_bi]

Hello, The password would be whatever you or your system admin set up at install. I just created a new PostrgreSQL connection and supplied the qlogs_reader username and password and used that new connection in the app. I think the main issue was the fact that we had multi node setup and the allowed IPs were not set up right.