Testnet observability: metrics, cold-path probe, health workflow - #436

Merged

aojea merged 10 commits into

google:mainfrom

aojea:observability-metrics

Sep 18, 2026

Commits on Sep 18, 2026

controlplane: export mesh state and per-route request metrics
aojea
committed
52fd80a
router: serve /metrics, /healthz and /readyz on --metrics-addr
aojea
committed
4a33506
node: export mesh-membership gauges and add --metrics-addr
aojea
committed
06494ec
deploy: scrape the control plane, routers and canaries with Managed Prometheus
aojea
committed
e9f4234
charts: metrics ports, HTTP probes and monitoring toggles
aojea
committed
fbb8774
deploy: retire the cop canary
aojea
committed
1581afa
controlplane: delete node records that outlived their session
aojea
committed
a4faefb
deploy: retire the agnhost canary
aojea
committed
0a16ee7
deploy: probe the cold path every 15 minutes and on every rollout
aojea
committed
448234a
ci: check each testnet's health every half hour
aojea
committed
a40a969