Percona Cluster and flow control

I am using Percona Cluster 5.5.39 and we have been having some issues with cluster flow control pauses. I think that this is a matter of a very high volume of transactions. One of the recommendations I have seen is to increase the fc_limit value. We are using Single master / multiple slaves and currently have the following gcs properties set (among others) “gcs.fc_limit=500; gcs.fc_master_slave=YES; gcs.fc_factor=1”. I was wondering if it is ok to increase fc_limit even further to something like 1000. Would there be any negative effects doing this that I should consider?

Also, we have had some other issues with large transactions, and we are breaking up those queries into smaller chunks, but we are concerned about the fact that we usually don’t realize these inserts or deletes are problematic until the data size of the affected tables grows to a breaking point. When this happens, it pauses all other inserts/deletes/updates sometimes for several minutes. I know that Galera 4.0 is supposed to handle large and long transactions. Any idea when this new version of Galera might be available in Percona XTraDB Cluster?

We don’t know yet when Codership is going to be releasing Galera 4.0 but in my own opinion you need to plan on upgrading to the latest PXC 5.6.x version as many fixes/improvements have been added since.