merge prometheus2 targets into prometheus1

I am not exactly sure what that means exactly, but prometheus1 needs to supplant prometheus2 so the latter can be retired and replaced with a replica that's highly available.

I think it means we need to deploy the prometheus-alerts repository onto the prometheus1 server. We might also need to do something about all those dashboards in Grafana.

We'll run both servers in parallel for a year (or whatever the prometheus2 retention period is) and then retire prometheus2.

As usual, notify users and all.

flight check:

• deploy prometheus-alerts repository on prom1 - currently does not produce alerts since prom1 doesn't scrape the same targets
• configure all of the scrape targets from prom2 onto prom1 and possibly firewall rules to make prom1 able to poll the metrics. create a long (1 year) silence for all alerts with team tags other than TPA. let prom1 gather some data for a little while and verify that scraping is working for all sources
• deploy dashboards that are used on grafana2 for the other teams onto grafana1 and check that it's showing mostly the same information
• copy over any other bit of prometheus and alertmanager configuration from prom2 to prom1 that's not currently common to both
• when we're confident enough create a long (2 years) silence on karma2 for all alerts. then remove the silence from prom1 to make that one send alerts instead
• after maybe two weeks of checking if everything behaves the same, create an issue for decommisioning prom2 after 1y