Hi Fan , andycutright , and Tim Schmolzi
Thanks for reporting this issue - turns out we’re trying to execute Prometheus 1.x (delete series) API calls against Prometheus 2.x backend. We’re tracking this issue in our JIRA under [url][PMM-2845] pmm-admin purge causes client to panic - Percona JIRA , and we are targeting our October release (tentatively Oct 10) for a fix.
[url]HTTP API | Prometheus
[url]HTTP API | Prometheus