Claus Ibsen created CAMEL-24272:
-----------------------------------
Summary: Use BackgroundTask instead of ForegroundTask for
reconnection loops so they are visible in internal tasks
Key: CAMEL-24272
URL: https://issues.apache.org/jira/browse/CAMEL-24272
Project: Camel
Issue Type: Improvement
Reporter: Claus Ibsen
Several components use ForegroundTask for reconnection/retry loops that run in
background threads. Because ForegroundTask blocks the calling thread and
completes quickly per iteration, these retries are not visible in the
internal-tasks dev console, CLI (camel get internal-task), or TUI Internal
Tasks tab.
Switching to BackgroundTask would register these retries with the
TaskManagerRegistry, letting operators observe reconnection attempts, retry
count, delay, elapsed time, and errors.
Components to update:
* *camel-google-pubsub* - GooglePubsubConsumer: PubSub async-pull reconnection
loop with 5s backoff runs in a background thread
* *camel-zookeeper* - ZooKeeperConsumer: reconnection backoff in background
OperationsExecutor thread
* *camel-salesforce* - SubscriptionHelper: CometD handshake retry and subscribe
retry delays in CometD listener background threads (2 usages)
* *camel-hazelcast* - HazelcastSedaConsumer: consumer error-recovery delay loop
in background run() thread
* *camel-mongodb-gridfs* - GridFsConsumer: entire consumer polling loop
implemented as unlimited ForegroundTask in a background executor (misuse of
ForegroundTask)
See also CAMEL-24271 for the same issue in camel-kafka.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
