o11y on the app server layer #121

New issue

Closed

opened 2026-07-20 19:19:51 +00:00 by coilysiren · 1 comment

coilysiren commented

2026-07-20 19:19:51 +00:00

Owner

The standard things

signoz
glitchtip (thats what I use right?)
off the self hosted product analytics app that I haven't decided on yet
...?

signoz carries a lot

The standard things - signoz - glitchtip (thats what I use right?) - off the self hosted product analytics app that I haven't decided on yet - ...? signoz carries a lot

coilyco-ops referenced this issue

2026-07-25 13:36:45 +00:00

[interactive] Resolve remaining Eco App decisions and live checkpoints #155

coilyco-ops added the

autonomy

async-consult

label

2026-07-25 15:00:18 +00:00

coilyco-ops referenced this issue

2026-07-26 00:11:18 +00:00

Grafana + Prometheus Exporter #11

coilyco-ops commented

2026-07-26 00:22:20 +00:00

Member

Decision: keep the app-server observability contract lean and use the fleet panes that already exist.

Logs - stdout and stderr are collected by the kai-server SigNoz filelog agent. Application and Discord-worker logs stay public-safe and structured enough for the shared ingest pipeline.
Errors - the existing Sentry-compatible SDK remains the code boundary. Infrastructure#395 owns the DSN cutover from hosted Sentry to GlitchTip. #44 owns the currently malformed eco-app DSN.
Metrics - Kubernetes and container health stay in VictoriaMetrics. Eco game and runtime metrics come from mods/telemetry, not a second app scraper.
Traces - no Python request tracing until a concrete latency question or SLO makes the added SDK and cardinality contract useful. Eco game traces remain under #127.
Product analytics - do not select or deploy a product-analytics system without named product questions, a retention period, and an event privacy contract. Page views alone do not justify another stateful service.

This resolves the pane split without new app code. #44, #127, and infrastructure#395 carry the remaining live configuration work. Closing the open-ended observability brainstorm.

Decision: keep the app-server observability contract lean and use the fleet panes that already exist. * **Logs** - stdout and stderr are collected by the kai-server SigNoz filelog agent. Application and Discord-worker logs stay public-safe and structured enough for the shared ingest pipeline. * **Errors** - the existing Sentry-compatible SDK remains the code boundary. Infrastructure#395 owns the DSN cutover from hosted Sentry to GlitchTip. #44 owns the currently malformed eco-app DSN. * **Metrics** - Kubernetes and container health stay in VictoriaMetrics. Eco game and runtime metrics come from `mods/telemetry`, not a second app scraper. * **Traces** - no Python request tracing until a concrete latency question or SLO makes the added SDK and cardinality contract useful. Eco game traces remain under #127. * **Product analytics** - do not select or deploy a product-analytics system without named product questions, a retention period, and an event privacy contract. Page views alone do not justify another stateful service. This resolves the pane split without new app code. #44, #127, and infrastructure#395 carry the remaining live configuration work. Closing the open-ended observability brainstorm.

coilyco-ops closed this issue

2026-07-26 00:22:20 +00:00

No milestone

No project

No assignees

2 participants

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

coilyco-gaming/eco-app#121

No description provided.

Rows
Columns