Uptime monitors: support cross-region POPs for all org data regions

Uptime monitors for US-region orgs only receive checks from US-based points of presence (POPs). Orgs on the DE data region similarly only receive checks from DE-based POPs. There is no way to get geographically distributed checks that span both US and EU nodes.

Current behavior

The uptime checker service (getsentry/uptime-checker) is deployed per Sentry data region with a fixed POP set (gocd pipeline config):

• us region → us-west-or, us-east-sc, us-east-va
• de region → de-west-de, de-west-nl, de (main)

The UPTIME_REGIONS setting in sentry (src/sentry/conf/server.py) controls which checker nodes are visible to a deployment. Cross-region nodes are not included.

Gap

Orgs whose users or infrastructure span multiple continents cannot get uptime checks from outside their own data region's POPs. For example, a US-region org with EU end users cannot verify reachability from European nodes without migrating their entire Sentry data region — a one-way, high-cost operation.

Options

• Cross-region POP access: Add remote DE POP endpoints to the US UPTIME_REGIONS config (and vice versa). UptimeRegionConfig already has an api_endpoint field (src/sentry/conf/types/uptime.py) suggesting this routing pattern was anticipated. The regionWarning component already surfaces a data-residency disclaimer for this case.
• Global POP expansion: Deploy additional checker POPs that are shared across all data regions rather than scoped per-region.
• User-selectable regions: Expose region selection in the uptime monitor form, allowing orgs to opt in to specific POPs (additive, not a replacement for the above).

Start with UptimeRegionConfig in src/sentry/conf/types/uptime.py and the UPTIME_REGIONS configuration in src/sentry/conf/server.py, then review the uptime-checker GoCD pipeline configuration. Compare the proposed cross-region, global-POP, and user-selectable approaches, including the existing regionWarning behavior. Done requires a decided scope and an implementation plan that preserves the stated data-residency expectations.