- job_name: node/cadvisor
honor_labels: true
honor_timestamps: true
scrape_interval: 10s
scrape_timeout: 10s
metrics_path: /metrics/cadvisor
scheme: https
kubernetes_sd_configs:
- role: node
namespaces:
names:
- <namespace>
tls_config:
insecure_skip_verify: true
- I use the container_memory_working_set_bytes metrics to retrieve the RAM used by my containers.
- When i redeploy my pod, i have 2 sample by promql 'instant query" (without the time parameter). Indeed, the sample related to the previous pod/deployment is still returned during few minutes,
- If i set honor_timestamps to false instead true, i have only the last sample active, it works.
Could you more information on this parameter?
Thanks
Loïc