Services: one copy down of several is degraded, not down #53

Merged
jcoffey-dev merged 1 commits from fix/service-redundant-parts into main 2026-11-16 06:19:10 +00:00
1 Commits
Author SHA1 Message Date
jcoffey-dev 2bba03db13 Services: one copy down of several is degraded, not down
ci / check (pull_request) Successful in 4m24s
ci / postgres (pull_request) Successful in 7m28s
A service's state was the worst of its parts, so one failed
redis-server among six cache hosts, or one exited app container among a
hundred, marked the whole service down: down on the service map, and
downtime in the availability report that never happened.

Now a unit or container a service has on several hosts is redundant: a
copy down while others still run makes the service degraded, and its
word says how many ("1 of 6 redis-server.service down"). Every copy
down, or only one, is down as before. Each copy's own alert is
unchanged. Documented in formats.md, topology.yaml.

Closes #52.
2026-11-15 22:59:49 -07:00