Monitoring: application telemetry in production (OpenTelemetry)

Aperta
#2,717 0 commenti 0 reazioni 0 assegnatari Vedi su GitHub

I maintainer di solito rispondono entro 3 giorni

Nessuno ha ancora preso questa issue.

Valutazione

Difficoltà
5/5
Tempo stimato
Più di una settimana
Idoneità per principianti
12/100
Tipo di issue
Funzionalità
Chiarezza
Abbastanza chiara
Stato di attività
Attiva
Stack tecnologico
apache, grafana, prometheus, python

Direzione di ricerca

This is a design issue, not a starter task: the checklist asks for decisions on overhead, trace contents, packaging and the collector setup before code. Read the context in #2716 and the dev prototype in #2484, then look at oeplatform/wsgi.py and podman/monitoring/, the places the issue names for the SDK setup and the new units. Done means the open items are answered in the thread, with overhead measured on a request path.

Scritto dal modello di indicizzazione a partire dal testo della issue.

Descrizione

deployment Infrastructure type: enhancement :gear:

Part of #2412. A later phase; not needed for the first monitoring version (#2716).

Why

#2716 watches the platform from outside: probes, host, containers, database, usage counts from the access log. It cannot say why a request is slow (which view, which queries). That needs telemetry from inside the application: traces and request metrics.

What is known

  • opentelemetry-instrument cannot instrument production. The production container runs Apache + mod_wsgi; the Python processes are mod_wsgi daemon processes, so a wrapper has nothing to start. The SDK must be set up in oeplatform/wsgi.py, once per daemon process.
  • The dev-side prototype is #2484, reduced to a development tool (see #2663). This issue is the production half.
  • The data path already exists: an OpenTelemetry Collector as another Quadlet unit beside podman/monitoring/, request metrics into the existing Prometheus, traces into Jaeger v2 or Grafana Tempo.

To decide and do

  • Overhead measured on the request path before anything reaches production (it runs on every request).
  • What a trace may contain: URL query strings can carry tokens and must be scrubbed; no user names.
  • Packages pinned in requirements-telemetry.txt; whether the production image installs them, or a build argument switches them on.
  • Collector and trace store as oep-mon-* units, localhost-only like the rest.
  • Relation to the planned slow-request log (one structured line per slow request): traces may replace it, or it comes first because it is cheaper.
Lingua principale
Python
Stelle
65
Fork
28
Merge medio
4g 7h
PR unite (30g)
25

Preparare l'ambiente

Apri in Codespaces

Avvia il container di sviluppo del progetto nel browser, con il tuo account GitHub.

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di OpenEnergyPlatform/oeplatform

Tutte le issue di OpenEnergyPlatform/oeplatform

Issue simili

Altre issue su Python

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.