Restart eco-server to validate tailnet loopback fix #34

New issue

Closed

opened 2026-06-17 07:45:23 +00:00 by coilysiren · 0 comments

coilysiren commented

2026-06-17 07:45:23 +00:00

Owner

Originally filed by @coilysiren on 2026-04-29T05:15:51Z

🤖 Filed by Claude Code on Kai's behalf.

Restart eco-server to pick up the new OtlpMetricsEndpoint and validate the tailnet loopback fix from coilysiren/infrastructure#71.

State at filing:

Live Configs/EcoTelemetry.json on kai-server is already pointing at http://localhost:30428/opentelemetry/api/v1/push (the new NodePort path). No further config edit needed.
vmsingle-nodeport Service is up in the observability namespace, NodePort 30428.
eco-server has not been restarted, so it's still running with the old in-memory exporter pointed at the dead tailnet IP.

To do:

coily eco restart
Check Logs/EcoTelemetry/smoke-probe.txt for a non-failing probe line.
Query vmsingle for eco_* series and confirm data is landing:
- curl -s 'http://<vmsingle-host>:8428/api/v1/label/__name__/values' | jq '.data | map(select(startswith("eco_")))'
- or browse Grafana at https://grafana.coilysiren.me.
If green: close this issue, close coilysiren/infrastructure#71, and close the loopback chapter of #5.
If still failing: re-open #71 with the new smoke-probe output - means the NodePort path is also blocked from host userspace, which would be a different bug.

Cross-links: coilysiren/infrastructure#71 (cluster-side fix), #5 (original metrics-not-arriving symptom).

🤖 Filed by Claude Code on Kai's behalf.

Migrated from coilyco-flight-deck/eco-telemetry#1 during the eco repo consolidation (coilysiren/inbox#100). Reinstated per Kai - this item survives the cycle and belongs in eco-app.

_Originally filed by @coilysiren on 2026-04-29T05:15:51Z_ > 🤖 Filed by Claude Code on Kai's behalf. Restart eco-server to pick up the new `OtlpMetricsEndpoint` and validate the tailnet loopback fix from coilysiren/infrastructure#71. **State at filing:** - Live `Configs/EcoTelemetry.json` on kai-server is already pointing at `http://localhost:30428/opentelemetry/api/v1/push` (the new NodePort path). No further config edit needed. - `vmsingle-nodeport` Service is up in the `observability` namespace, NodePort 30428. - eco-server has *not* been restarted, so it's still running with the old in-memory exporter pointed at the dead tailnet IP. **To do:** 1. `coily eco restart` 2. Check `Logs/EcoTelemetry/smoke-probe.txt` for a non-failing probe line. 3. Query vmsingle for `eco_*` series and confirm data is landing: - `curl -s 'http://<vmsingle-host>:8428/api/v1/label/__name__/values' | jq '.data | map(select(startswith("eco_")))'` - or browse Grafana at <https://grafana.coilysiren.me>. 4. If green: close this issue, close coilysiren/infrastructure#71, and close the loopback chapter of #5. 5. If still failing: re-open #71 with the new smoke-probe output - means the NodePort path is also blocked from host userspace, which would be a different bug. Cross-links: coilysiren/infrastructure#71 (cluster-side fix), #5 (original metrics-not-arriving symptom). > 🤖 Filed by Claude Code on Kai's behalf. --- _Migrated from coilyco-flight-deck/eco-telemetry#1 during the eco repo consolidation (coilysiren/inbox#100). Reinstated per Kai - this item survives the cycle and belongs in eco-app._

coilysiren added the

priority

label

2026-06-17 07:45:23 +00:00

coilysiren closed this issue

2026-07-06 15:52:37 +00:00