Azure/azure-event-hubs-node

run long-running scenario tests without issues

joshgav opened this issue · 4 comments

joshgav commented
  1. ClientSenderSendReceive - Generic send and receive on client.
  2. HighThroughput - Runs on a 100TU namespace and saturates ingress.
  3. HighThroughputPartitionSender - Runs on a 100TU namespace and saturates ingress with partition senders.
  4. EphShuffleHostsLargeShards - Registers and unregisters hosts as part of the regular iteration to introduce excessive partition moves. No sends in this scenario.
  5. EphLargeShards - Runs EPH on 256 partition entity.
amarzavery commented

HighThroughput - Runs on a 100TU namespace and saturates ingress.

screen shot 2018-08-08 at 7 21 50 pm

amarzavery commented

ClientSenderSendReceive - Generic send and receive on client.

screen shot 2018-08-08 at 7 26 25 pm

amarzavery commented

HighThroughputPartitionSender - Runs on a 100TU namespace and saturates ingress with partition senders.

screen shot 2018-08-12 at 2 26 24 pm

amarzavery commented

EphShuffleHostsLargeShards - Registers and unregisters hosts as part of the regular iteration to introduce excessive partition moves. No sends in this scenario.

Ran the test by starting with 3 EPH instances. It was able to balance the load effectively across 256 partitions. Later brought 1 EPH instance down and the load was balanced effectively again.