[08:56:03] <Krinkle>	 ori: not that I know, no.
[08:56:48] <Krinkle>	 there are numerous metrics per instance / per pod in prometheus that might let you plot how it evolves typically over a pods lifetime after a deployment/restart.
[08:57:25] <Krinkle>	 however latency is collected with mtail centrally on centrallog#### so the {instance} label for mediawiki_duration won't give you that
[09:07:46] <Krinkle>	 other metrics do come through, maybe latency as well. for example, apcu data. `sum(php_apcu_memory{type="total", cluster="", kubernetes_pod_name=~"mw-web.*", kubernetes_pod_name!~".*canary.*"}) by (kubernetes_pod_name)` (shows 14,000 metrics; times out; probably includes past pods, so might be worth using a templte variable limited to current ones by uptime and then showing "All" of those only).
[09:08:01] <Krinkle>	 https://grafana-rw.wikimedia.org/d/000000327/apache-fcgi
[09:08:13] <Krinkle>	 https://grafana-rw.wikimedia.org/d/35WSHOjVk/application-servers-red-k8s
[09:08:42] <Krinkle>	 https://codesearch.wmcloud.org/search/?q=mediawiki_http_requests_duration (see mtail file)