We are running kong within Kubernetes and have seen this intermittent critical error within the gateway logs that leads to a 499 error on the respective request.
*[crit] 35#0: 1970313 [lua] balancer.lua:533: could not obtain list of upstreams: [postgres] receive_message: failed to get type: timeout
We are running kong 2.5.0 as a daemonset on Azure AKS using postgres for the kong configuration.
This has been occurring every few hours as single errors, except for an incident earlier in the week where we seen a burst of these errors over a 4 hour period that swamped postgres with new connection requests.
Has anybody seen this error before or have an insight as to what the root cause could be?