TCP connections pileup - Apigee Edge

Hi,

We have bunch of applications of a system supported by Apigee edgemicro as the gateway.

There are 5 microgateway (edgemicro) servers supporting this.

Max connections:3000

Timeout : 15 sec

On a peak day, there was a sudden spike in the incoming traffic due to which connections spiked up to almost the max value in all servers. After the spike got subsided, one of the microgateway (edgemicro) server got recovered immediately by freeing up connections and it came back normal (around 300 connections range)

The other servers did not free up connections and were still in the range of 2000-3000 for a longer time which resulted in many 504's. Only after the restart of the edgemicro processes, they got recovered to the normal range of connections.

Question:

Are there any factors on the server (either at OS level or at the edgemicro node process level) that would impact this connections lingering around ?

Are there chances of connections becoming stale and recover only after restart?

1 0 198
0 REPLIES 0