APIGEE Edge UI does not show Analytics Data for Proxy

Not applicable

We are currently at OPDK Version 4.15.07.00.

For the past week , Analytics data is not reflecting on the Edge UI. However by enabling the Trace, the flow is logged as expected.

On analysis of the /opt/apigee4/var/log/qpid/qpidd.log,the below warning entries are logged;

2016-07-25 09:15:16 warning Journal "ax-q-axgroup001-consumer-group-001": Enqueue capacity threshold exceeded on queue "ax-q-axgroup001-consumer-group-001"

2016-07-25 09:15:16 warning Exchange ax-ex-axgroup001 cannot deliver to queue ax-q-axgroup001-consumer-group-001:
Enqueue capacity threshold exceeded on queue "ax-q-axgroup001-consumer-group-001". (JournalImpl.cpp:597)

Ran the pruning of Analytics with the Script provided in the Operations Guide. It reduced the Disk space usage from 88% to 76% but original issues still remains.

opt/apigee4/bin/qpid-stat.sh -q shows around 9.6K as message-in and 56 as message-out.

Would request help with the defined mechanism to flush the QPID queue for OPDK Version 4.15.07.00

Thanks!

0 5 832
5 REPLIES 5

This indicates there is probably a problem with the process that drains the Qpid queue into the analytics database. The first step is to check each of the processes involved in the analytics backend. Ensure that the qpid-server, Qpid, and PostgreSQL services are all up and running. The qpid-server process is responsible for pulling messages off the queue and inserting them into the PGSQL database. Since you are able to query your queue status and you already purged data from PGSQL, it is likely that the qpid-server process is the culprit here. Ensure that it is running, restart it if necessary, and check system.log for errors.

Not applicable

Thanks for the response Paul. Yes- I did check the components on both the Nodes involving Analytics ( We have a 5 node set up ) and they look good ( all-status.sh and check.sh- All processes and services are running fine with a 200 response). Also restarted all components on both the Nodes involving Analytics. The logs are still flooded with - Enqueue capacity threshold exceeded on queue.

Not applicable

@Ranjit Have you seen this community article? Resize the qpid queues This gives details on how to resize the qpid queues to resolve this error

Thanks for the response Stuart. Yes, I had read that and as it had instructions for a specific OPDK, we opened a support ticket . There are slight modifications required and we are still working on the steps with support from Apigee. Once successfully completed, will paste the details of steps we carried out.

Not applicable

The above reported issue was resolved with the help of APIGEE Support. Detailing the resolution below ->

The servers on which the Analytics components( QPID,QPID Server, Postgress, Postgress Server) were hosted ran out of disk space. This triggered the Postgress servers (Master and Secondary) to be out of sync.

select max(client_received_start_timestamp) from analytics."your_org_name.public.fact" returned different entries from both servers.

Resolution was to bring the Postgress servers in Sync using the below scripts. (

/opt/apigee4/bin/apigee-postgres-replication-setup.sh,

postgres-check-standby.sh,

postgres-check-master.sh

)