[
https://issues.apache.org/jira/browse/CAMEL-24644?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18113312#comment-18113312
]
Claus Ibsen commented on CAMEL-24644:
-------------------------------------
Merged to main via https://github.com/apache/camel/pull/26191 (commit
fae3442189680a977a9e8bcdef8fd06c17fa0766). Will be in 4.23.0.
_Claude Code on behalf of davsclaus_
> camel-opa - add a producer health check for the OPA server
> ----------------------------------------------------------
>
> Key: CAMEL-24644
> URL: https://issues.apache.org/jira/browse/CAMEL-24644
> Project: Camel
> Issue Type: Improvement
> Reporter: Andrea Cosentino
> Assignee: Andrea Cosentino
> Priority: Minor
> Fix For: 4.23.0
>
>
> h2. Problem
> {{camel-opa}} (CAMEL-24634) fails closed by design: when the policy cannot be
> evaluated at all - OPA unreachable, timed out, or answering with an error -
> the producer throws {{OpaPolicyEvaluationException}} and
> {{OpaSecurityPolicy}} throws {{CamelAuthorizationException}}. That is the
> correct behaviour for an authorization component, but it means an OPA server
> that is down takes down *every* exchange through the route.
> There is currently no way to observe that before traffic starts failing. The
> only signal is the failures themselves.
> h2. Proposal
> Register a producer health check that probes the OPA server's {{/health}}
> endpoint, so an unreachable policy decision point shows up in the health
> registry (and therefore in {{/q/health}} and the readiness probe) rather than
> only in the error logs.
> Existing precedent for a component-supplied health check:
> * {{RemoteFileProducerHealthCheck}} (camel-ftp-common)
> * {{KafkaProducerHealthCheck}} (camel-kafka)
> * {{PQCStatefulKeyHealthCheck}} (camel-pqc)
> Notes:
> * {{/health}} is the endpoint the {{camel-test-infra-opa}} container service
> already uses as its Testcontainers wait strategy, so the probe is known to
> work.
> * The check should report DOWN with the underlying reason, not swallow it,
> since the whole point is telling an unreachable PDP apart from a policy that
> denied.
> * Worth considering whether the check belongs on {{OpaSecurityPolicy}} too -
> it has the same exposure but no endpoint to hang a producer health check on.
> h2. Scope
> {{main}} only. Additive.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
