Exercise 2: Collapsing 15 Dashboards Into One via a Template Variable — Possible Solution
====================================================================

Panel query, using a $service template variable:

    sum(rate(http_requests_total{job="$service"}[5m]))

-- How a template variable replaces 15 separate dashboards --
--
-- Right now, the team has effectively hand-built the SAME dashboard
-- 15 times, once per microservice, differing only in which specific
-- job/service name each dashboard's queries happen to hardcode. A
-- $service template variable replaces that hardcoded value with a
-- dropdown at the top of a single dashboard, populated by a
-- label_values() query against Prometheus's own job label (the exact
-- label obs1-3's own scrape_configs automatically attaches to every
-- metric). Every panel's query -- request rate, error rate, latency,
-- whatever the original 15 dashboards each repeated -- gets written
-- ONCE, referencing $service instead of any one specific service
-- name. Selecting a different value from the dropdown re-runs every
-- panel's query with that new value substituted in, showing the same
-- dashboard layout for whichever service was picked, with no need to
-- maintain 15 separate copies that would otherwise all have to be
-- updated in lockstep every time a panel changes.

WHY THIS WORKS AS AN ANSWER
------------------------------
This writes a correctly parameterized panel query using $service in
place of a hardcoded job name, then explains concretely how the
existing 15-dashboard duplication collapses into one dashboard plus a
dropdown, directly tying the fix back to the job label obs1-3
established scrape_configs attaches automatically.