fix(services): считать failover по IP Health а не по строке ноды
quality / commitlint (push) Skipped
quality / changes (push) Successful in 9s
quality / docker-check (push) Skipped
CD / update-wiki (push) Successful in 4s
quality / web (push) Successful in 52s
quality / api (push) Successful in 42s
CD / quality (push) Successful in 1m50s
CD / publish (push) Successful in 1m43s

Панель брала service_nodes, которую затирал group apply.
Теперь тот же binding ip_health, что таблица активов; group не пишет в ноду.

Co-authored-by: Cursor <[email protected]>
This commit is contained in:
Denozordec
2026-08-20 15:08:10 +07:00
co-authored by Cursor
parent 994e79e118
commit 7938d2f707
7 changed files with 161 additions and 75 deletions
@@ -386,7 +386,8 @@ function applyAggregatedStatus(
{ colo, provider: statusProvider },
);
const matchedNode = repos.findNodeByIp(db, target.ip);
if (matchedNode && matchedNode.enabled) {
// Binding-scope only: group apply must not clobber node with its own fetch failed.
if (matchedNode && matchedNode.enabled && target.scope === "binding") {
repos.updateNode(db, matchedNode.id, {
health_status: node,
consecutive_failures: failures,