🚀(staging) don't default to 3 Gunicorn workers

Currently we let SATOSA store some state in-memory. This cannot
work if we have multiple workers.

docker/files/usr/local/etc/gunicorn/satosa.py

@@ -5,7 +5,6 @@ python_path = "/app"
 # Run
 graceful_timeout = 90
 timeout = 90
-workers = 3
 
 # Logging
 # Using '-' for the access log file makes gunicorn log accesses to stdout

src/helm/env.d/dev/values.oidc2fer.yaml.gotmpl

@@ -7,6 +7,7 @@ satosa:
   replicas: 1
   envVars:
     BASE_URL: https://oidc2fer.127.0.0.1.nip.io
+    GUNICORN_CMD_ARGS: --workers=1
     CLIENT_DB_JSON: |-
       {
         "oidc-test-client": {

src/helm/env.d/staging/values.oidc2fer.yaml.gotmpl

@@ -7,6 +7,7 @@ satosa:
   replicas: 1
   envVars:
     BASE_URL: https://oidc2fer-staging.beta.numerique.gouv.fr
+    GUNICORN_CMD_ARGS: --workers=1
     SAML2_BACKEND_CERT: {{ .Values.SAML2_BACKEND_CERT | toYaml | indent 8 }}
     SAML2_BACKEND_KEY: {{ .Values.SAML2_BACKEND_KEY | toYaml | indent 8 }}
     OIDC_FRONTEND_KEY: {{ .Values.OIDC_FRONTEND_KEY | toYaml | indent 8 }}