mirror of
https://github.com/luxfi/monitoring.git
synced 2026-10-30 15:24:48 +00:00
zeekayandHanzo Dev
3e48e928f9
luxd v1.34 serves the JSON-RPC/info/health APIs under /v1/* not /ext/*, so the validator-health exporter was polling dead paths and every mainnet validator falsely reported down. Repoint exporter.py to /v1/*. Add the incident-critical node/pod memory pipeline on lux-k8s (the owner is debugging luxd OOM-evictions): - node-exporter DaemonSet (per-node host memory, all nodes incl luxd nodes) - kube-state-metrics (pod phase/restarts/OOM+Evicted reasons, node conditions) - vmagent RBAC + scrape jobs for node-exporter, kubelet cAdvisor (pod mem, drop-filtered), kube-state-metrics; heavy pods pinned to roomy non-luxd nodes - Grafana dashboard 'Lux — Cluster Nodes & OOM' (uid lux-nodes-oom) Read-only w.r.t. luxd/ns lux-mainnet. Ships via the existing vmagent -> metrics.hanzo.ai hub. Co-authored-by: Hanzo Dev <[email protected]>