While running a low-volume load test on the ESN, one node, play, did not maintain sync with the rest of the network. It's transaction count remained at 15072, while the rest of the network was posting new transactions at a rate of one per minute. At 20:51:23 UTC, when the ledger transaction count on the rest of the pool reached 15175, the load generator was killed. Immediately at this time, play began to resync with the rest of the pool. This resync was successfully completed at 20:53:01.
The issue to investigate is why was play unable to obtain and maintain synchronization with the remainder of the pool while the load generator was running an a low rate?