We are running timescaledb single helm chart- helm-charts/charts/timescaledb-single at main · timescale/helm-charts · GitHub on Google Kubernetes Engine (GKE)
We’re seeing a connection reset by timescale and then a reuse of the same dead connection (which is likely in our code) - something is not correctly set with a dropped connection in timescaledb
Application is disconnecting but that disconnect is as expected from timescale
When it times out on some non-active connection
We are experiencing disconnection issues from our applications running externally outside GKE and seeing logs like here
2022-12-13 01:15:36 UTC [634465]: [6397ca0b.9ae61-5] postgres@test-reproducer-1,app=[unknown] [00000] LOG: disconnection: session time: 0:34:52.827 user=postgres database=test-reproducer-1 host=10.92.2.1 port=27603
2022-12-13 01:15:36 UTC [636983]: [6397cdd2.9b837-5] postgres@test-reproducer-1,app=[unknown] [00000] LOG: disconnection: session time: 0:18:46.106 user=postgres database=test-reproducer-1 host=10.128.15.222 port=4892
-
What does this disconnection session log mean? is our service connecting at all?
-
why are we experiencing dead connections from timescale- why?