Karl Stoney

I should add by the way that sometimes they are getting recorded; it seems to specifically be when we get the error `upstream reset: reset reason: connection termination`, that they're...

@keithmattix - probably best to leave this one with me actually, it seems hit and miss for me so there's got to be a correlating factor that I'm missing. Will...

+1 for this, I have a situation where I effectively want to disable tcp keep alive

@melloc ``` server.server.on('connection', (socket) => { socket.setTimeout(timeout); }); ```

Sorry I never got back to you here, completely missed the question. > hmmm, I believe in normal case the span should be created and reported. Is your request a...

This caught me out again today and caused me to rabbit hole, investigating an incident from the traces i jumped to the conclusion the issue was in zone-a, but that...

Extremely hacky way i worked around this: Create a custom `istio-pilot` image with a wrapper for `pilot-agent`: ``` #!/bin/bash METADATA_SERVER="http://metadata.google.internal/computeMetadata/v1" ZONE_INFO_PATH="instance/zone" NODE_NAME_PATH="instance/name" ZONE=$(curl -f --max-time 2 "${METADATA_SERVER}/${ZONE_INFO_PATH}" -H "Metadata-Flavor: Google"...

@kyessenov or @lei-tang is there any plan to fix this? I'm unable to use distroless images because of it, as i have to run a funky init script to set...

Seeing as there are several things like this since telemetryv2, is it correct to continue to mark things like “client and server telemetry reporting” and “traffic control” as “Stable” on...

Similar problem but more serious manifestation is when you get DC errors (source service disconnecting, for example - because it's hit a read timeout) you don't get metric labels either:...