Currently, there is configuration server.retention which is optional and specifies how many days back to keep in Prometheus storage.
However, since Prometheus 2.7, there is an additional parameter --storage.tsdb.retention.size controlling the size of the data retained. It is useful, as overprovisioning PVC is not always viable and sometimes we prefer losing data, rather than being out of storage.
Prometheus-operator HELM chart already supports retentionSize parameter, and it can be used together with retention. https://github.com/helm/charts/blob/f019fd53a81e7124df65be0d4da620795c251a2e/stable/prometheus-operator/templates/prometheus/prometheus.yaml#L84
It would be useful to have the same functionality in the standard Prometheus chart.
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or unsubscribe.