How to Catch and Handle DMN Errors in Camunda?

Camunda 7 Topics Discussion & Questions
, ,
1

Bonjour,

Est-ce que quelqu’un a déjà mis en place la capture des erreurs provenant de DMN dans Camunda via un gestionnaire d’erreurs (Error Boundary Event, Error End Event ou autre mécanisme) ?

Si oui, quelle configuration avez-vous utilisée côté BPMN/DMN pour propager et intercepter correctement ces erreurs ?

Je suis particulièrement intéressé par les retours d’expérience concernant la gestion des erreurs DMN et les paramètres nécessaires pour qu’elles soient capturées par le processus.

Merci d’avance pour vos retours.

========== Camunda ==========

camunda:
bpm:
auto-deployment-enabled: false
database:
type: postgres
schema-update: true
history-level: audit
datasource:
driver-class-name: ${DB_DRIVER:org.postgresql.Driver}
url: ${DB_URL:jdbc:postgresql://127.0.0.1:5431/nspbpmdb}
username: ${DB_USERNAME:postgres}
password: ${DB_PASSWORD:root}
hikari:
pool-name: CAMUNDA-NSP-HIKARI-POOL
max-lifetime: 600000
idle-timeout: 120000
keepalive-time: 60000
connection-timeout: 30000
maximum-pool-size: 10
minimum-idle: 2
connection-test-query: SELECT 1
validation-timeout: 3000
job-executor:
enabled: false
wait-time-in-millis: 500
max-jobs-per-acquisition: 16
lock-time-in-millis: 1200000
admin-user:
id: ${CAMUNDA_ADMIN_USER:admin}
password: ${CAMUNDA_ADMIN_PASSWORD:admin}
firstName: Admin
lastName: User

========== ORD Engine ==========

ord:
bascule:
default-delay-seconds: 45
max-delay-seconds: 7200
active: ${ORD_ACTIVE:false}
engine:
datasource:
hikari:
pool-name: CAMUNDA-ORD-HIKARI-POOL
max-lifetime: 600000
idle-timeout: 120000
keepalive-time: 60000
connection-timeout: 30000
maximum-pool-size: 5
minimum-idle: 1
connection-test-query: SELECT 1
validation-timeout: 3000
job-executor:
wait-time-in-millis: 5000
max-jobs-per-acquisition: 8
lock-time-in-millis: 1200000

2

I don’t think that the DMN engine throws any BPMN errors, so I’m not sure how you would catch them.
Remember that BPMN errors are to capture the expectable flow issues.

3

Hi,

Thanks for your reply.

I know that, by default, a DMN evaluation does not throw a BPMN Error that can be caught by an Error Boundary Event or an Error End Event.

What I’m actually looking for is whether anyone has implemented a workaround or a custom solution to handle DMN failures. For example:

  • Using an execution/task listener after the Business Rule Task.

  • Wrapping the DMN evaluation in a JavaDelegate and converting exceptions into BPMN Errors.

  • Using parse listeners or engine plugins.

  • Any other mechanism to detect unexpected DMN evaluation failures and route the process to an error-handling path…

I’m interested in real-world experiences rather than the standard behavior documented by Camunda.