Andrea Cosentino created CAMEL-24736:
----------------------------------------
Summary: camel-opa - OpaSecurityPolicy has no health check for the
OPA server
Key: CAMEL-24736
URL: https://issues.apache.org/jira/browse/CAMEL-24736
Project: Camel
Issue Type: Improvement
Reporter: Andrea Cosentino
Assignee: Andrea Cosentino
h2. Problem
CAMEL-24644 added a producer health check so an unreachable OPA server is
visible before it fails traffic. It covers the {{opa:}} producer only -
{{OpaSecurityPolicy}} has none:
{noformat}
$ grep -rn "HealthCheck" components/camel-opa/src/main/java/.../opa/security/
(no match)
{noformat}
That is the wrong way round. Both paths fail closed when OPA is unreachable,
but the security policy is the stricter of the two: a producer merely records a
deny verdict the route can inspect, while {{OpaSecurityPolicy}} throws
{{CamelAuthorizationException}} and stops the exchange outright. The path that
hard-fails every message has no health signal at all.
This was raised as an open question in CAMEL-24644 ("worth considering whether
the check belongs on OpaSecurityPolicy too - it has the same exposure but no
endpoint to hang a producer health check on") and never resolved.
h2. Why it is not just a copy-paste
{{OpaProducer}} registers into the {{producers}} repository because it *is* a
producer. {{OpaSecurityPolicy}} is a bean wrapped around a route segment; it
has no endpoint and no producer. It needs either registration into a different
repository or a small {{HealthCheckRepository}} of its own, keyed so that two
policies pointing at different servers stay distinct - the same id-collision
trap CAMEL-24717 had to fix, and with the same requirement that the id not
carry {{bearerToken}}.
h2. Scope
{{main}} only. Additive.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
