Question
Vodafone Espana SA
ES
Last activity: 3 Oct 2025 17:22 EDT
HFIX-C1166 High count of java.util.concurrent.TimeoutExceptions for Kafka
Hello team,
We have Upgraded Pega Platform from 8.6.6 to 23.1.4 and when we scan the system, this Critical Missing Hotfix appears:
HFIX-C1166 High count of java.util.concurrent.TimeoutExceptions for Kafka but we haven't received any CAD advise or INC.
After searching the description in MSP, I founded this explanation:
"A very high count of java.util.concurrent.TimeoutExceptions was observed after migration to external Kafka for running the stream service."
This issue primarily occurred during lag calculation with external streams or in environments with a large number of queue processors. To mitigate this, the Kafka AdminClient connections will be cached to reduce the number of connections to the external stream node, which helps avoid timeout exceptions.
We are not using external Kafka yet, so we understand that it is not necessary to install this Hotfix at the moment.
Please, could you help us to identify if this HF applies to us or is it advisable to install it even if we do not use the external kafka?
How to test it after install the HF?
Regards.
HCA Healthcare
US
install HFIX-C1166 only if you use external Kafka (or plan to soon) or you’re seeing TimeoutExceptions during lag checks; otherwise you can skip it. The fix caches Kafka AdminClient connections, which matters only with external Kafka or very large numbers of queue processors. If you run Pega’s built-in Stream service (no external Kafka endpoints, no Kafka auth/profile), the hotfix won’t change behavior. To double-check applicability, verify Stream > Settings shows external Kafka brokers, or look for DSS like decision/stream/* or messaging/externalkafka in use. If you do install, test by: clearing caches, restarting stream nodes, then exercising queue processors and real-time data flows; watch PDC/Logs for TimeoutExceptions drop to zero, confirm queue-processor lag stabilizes, and check Stream landing page health. Also validate AdminClient connection counts don’t spike under load and throughput/latency remains steady. Keep a rollback plan (snapshot) and note the hotfix ID in your release notes.