david-sykora

+1 we observe similar behavior on our servers

Also I found out two things: - it doesn't matter if you duplicate existing connection or create new one, it has no impact on leakage size: ```javascript for (var i...

Looks like it just takes some little time until GC swallows unused heap data. Btw. same things happens for standard redis connection

My bad. This is probably due to my misunderstanding of how Libvirt networking works. [Networks in "veryprivate" mode have all communication with "guest" disabled](https://libvirt.org/formatnetwork.html#network-config-with-no-gateway-addresses). I am trying to achieve the...

@anthosz Thank you for the pull request, it looks like a much more efficient solution than what I've been using ([HAProxy setup](https://blogs.oracle.com/cloud-infrastructure/post/deploying-highly-available-redis-replication-with-haproxy-on-oracle-cloud-infrastructure) - relatively resource demanding). I tried to test...

I used the workaround mentioned in the issue, the normal service works, but the master doesn't match any pods because the **isRedis** label is not assigned. I am attaching the...