Gostatsd with Graphite backend send log to stderr

[barrowkwan]

I am not sure what option to make the log not to send to stderr. Right now even INFO log will send to stderr.

Here is our configuration:

server-mode = 'standalone' # input is aggregated and sent to backend
ignore-host = true
metrics-addr = ':8125'
backends = ['graphite']
flush-interval = '10s'
max-readers = 8
max-parsers = 8
max-workers = 8
max-concurrent-events = 1024
max-queue-size = 100000
bad-lines-per-minute = 100 # don't put too much space for trash metrics
expiry-interval='1h'
log-raw-metric = true
[graphite]
address = "carbon-relay-ng.monitoring:2033"
dial_timeout = '15s'
write_timeout = '30s'
mode = 'legacy' 
global_suffix = ''

Here is the log in GCP, in textPayload the message severity is level=info

{
insertId: "xxxxxxxxxxxx"
labels: {
compute.googleapis.com/resource_name: "gke-xxxx-s-1f2202f0-xsp3"
k8s-pod/app: "gostatsd"
k8s-pod/pod-template-hash: "78c89b6648"
}
logName: "projects/xxxx-01/logs/stderr"
receiveTimestamp: "2021-12-03T22:24:29.492892849Z"
resource: {
labels: {
cluster_name: "xxxxxxx-apps"
container_name: "gostatsd"
location: "asia-southeast1"
namespace_name: "monitoring"
pod_name: "gostatsd-78c89b6648-qdrg8"
project_id: "xxxx-01"
}
type: "k8s_container"
}
severity: "ERROR"
textPayload: "time="2021-12-03T22:24:27Z" level=info msg="[{timer, myapp.api.ping.GET.latency, 0.055000, , } {counter, myapp.api.ping.GET.status.200, 1.000000, , } {timer, myapp.api_aggregate.latency, 0.055000, , } {counter, myapp.api_aggregate.status.200, 1.000000, , }]"
"
timestamp: "2021-12-03T22:24:27.885278648Z"
}

[barrowkwan]

Is it because of the log-raw-metric option?

[tiedotguy]

Yep - that's intended for debugging, it's basically logging the decoded udp packet, not the aggregated data on flush.

[barrowkwan]

@tiedotguy thanks!