davsclaus opened a new issue, #3261: URL: https://github.com/apache/camel-kamelets/issues/3261
## Citrus tests for the Kamelets that have none **148 of the 221 Kamelets** on `main` are used by no Citrus test under `tests/camel-kamelets-itest`: neither as `kamelet:<name>` nor in a Pipe. (The count leaves out `mqtt-source` and `mqtt-sink`, which #3260 removes.) Only 20 Kamelets carry `camel.apache.org/kamelet.verified: "true"`. Untested Kamelets do break without anyone noticing. The overhaul for 4.24 found Kamelets that could not start, options that never reached the component, and templates that changed the data (#3156, #3157, #3158). A test that runs the Kamelet would have caught each one. ### What a test should show At least one message through the Kamelet against a real or emulated service, and a check of what comes out: - **Sources:** the body and the headers they document. - **Sinks:** what arrives at the service. - **Actions:** the body and headers before and after. Where a Kamelet has an option with a special case, such as a password with `&`, a file name with spaces, or `getObject`, the test covers that case too. A Kamelet that passes gets `camel.apache.org/kamelet.verified: "true"`. ### What can run them Citrus 5.2 starts LocalStack/Floci, Kafka, Redpanda, MongoDB and PostgreSQL, plus any other image through `testcontainers: start: container:`. It can also stand in for an HTTP service (`http: server:`). So most families can be tested in CI: | Family | Untested | How | |---|---|---| | API | 2: `fhir-sink`, `fhir-source` | An HTTP mock where the Kamelet takes a base URL; otherwise manual | | AWS | 21: `aws-bedrock-agent-runtime-sink`, `aws-bedrock-agent-sink`, `aws-bedrock-text-sink`, `aws-cloudtrail-source`, `aws-cloudwatch-sink`, `aws-comprehend-sink`, `aws-ec2-sink`, `aws-kinesis-firehose-sink`, `aws-lambda-sink`, `aws-polly-sink`, `aws-redshift-sink`, `aws-redshift-source`, `aws-s3-event-based-source`, `aws-s3-streaming-upload-sink`, `aws-secrets-manager-sink`, `aws-ses-sink`, `aws-sns-fifo-sink`, `aws-sqs-batch-sink`, `aws-sqs-fifo-sink`, `aws-timestream-query-sink`, `aws-translate-action` | Floci (S3, SQS, SNS, Kinesis, DynamoDB, Lambda, SES, CloudWatch, EventBridge, Secrets Manager...). Bedrock, Comprehend, Polly, Translate, Redshift and Timestream may not be emulated: an HTTP mock, or manual | | Azure | 17: `azure-cosmosdb-sink`, `azure-cosmosdb-source`, `azure-eventhubs-sink`, `azure-eventhubs-source`, `azure-functions-sink`, `azure-servicebus-sink`, `azure-servicebus-source`, `azure-storage-blob-append-sink`, `azure-storage-blob-event-based-source`, `azure-storage-blob-sink`, `azure-storage-blob-source`, `azure-storage-datalake-sink`, `azure-storage-datalake-source`, `azure-storage-files-sink`, `azure-storage-files-source`, `azure-storage-queue-sink`, `azure-storage-queue-source` | Floci for Azure (`floci-az`), Azurite for storage | | Cache | 3: `infinispan-sink`, `infinispan-source`, `redis-source` | Infinispan and Redis containers | | Cloud | 10: `ceph-event-based-source`, `ceph-sink`, `ceph-source`, `github-commit-source`, `github-event-source`, `github-pullrequest-comment-source`, `github-pullrequest-source`, `github-tag-source`, `minio-sink`, `minio-source` | Floci S3 for Ceph and MinIO, a MinIO container; the GitHub sources: an HTTP mock | | Database | 12: `databricks-sink`, `databricks-source`, `mariadb-sink`, `mariadb-source`, `mysql-sink`, `mysql-source`, `oracle-database-sink`, `oracle-database-source`, `snowflake-sink`, `snowflake-source`, `sqlserver-sink`, `sqlserver-source` | MariaDB, MySQL, SQL Server and Oracle Free containers; Snowflake and Databricks: manual | | EIP | 3: `drop-header-action`, `drop-headers-action`, `header-matches-filter-action` | No service needed | | File | 10: `file-watch-source`, `ftp-sink`, `ftp-source`, `ftps-sink`, `ftps-source`, `scp-sink`, `sftp-sink`, `sftp-source`, `ssh-sink`, `ssh-source` | SFTP, FTP and SSH containers, and the local file system | | GCP | 11: `google-bigquery-sink`, `google-calendar-source`, `google-mail-source`, `google-pubsub-sink`, `google-pubsub-source`, `google-sheets-sink`, `google-sheets-source`, `google-storage-event-based-source`, `google-storage-sink`, `google-storage-source`, `google-vertexai-sink` | Floci for GCP (`floci-gcp`: Pub/Sub, Cloud Storage, BigQuery). The Google Workspace sources and Vertex AI: manual | | Geo | 2: `nominatim-geocode-action`, `ogcapi-features-action` | An HTTP mock where the Kamelet takes a base URL; otherwise manual | | Graphql | 1: `graphql-sink` | An HTTP mock where the Kamelet takes a base URL; otherwise manual | | IBM | 3: `ibm-cos-sink`, `ibm-cos-source`, `ibm-watson-language-sink` | Floci S3 for COS; Watson: manual | | JIRA | 3: `jira-oauth-source`, `jira-transition-issue-sink`, `jira-update-issue-sink` | An HTTP mock where the Kamelet takes a base URL; otherwise manual | | JSON | 3: `json-schema-validator-action`, `json-serialize-action`, `jsonata-action` | No service needed | | Kafka | 7: `is-tombstone-filter-action`, `kafka-batch-manual-commit-action`, `kafka-batch-source`, `kafka-manual-commit-action`, `message-timestamp-router-action`, `set-kafka-key-action`, `value-to-key-action` | Redpanda | | Mail | 2: `mail-imap-source`, `ms-exchange-online-imap-oauth-source` | GreenMail container | | Messaging | 12: `jms-amqp-10-sink`, `jms-amqp-10-source`, `jms-apache-artemis-sink`, `jms-apache-artemis-source`, `jms-ibm-mq-sink`, `jms-ibm-mq-source`, `pulsar-sink`, `pulsar-source`, `smpp-sink`, `smpp-source`, `spring-rabbitmq-sink`, `spring-rabbitmq-source` | Artemis, IBM MQ, RabbitMQ, Pulsar and SMPP containers | | Monitoring | 2: `snmp-source`, `splunk-hec-sink` | An HTTP mock where the Kamelet takes a base URL; otherwise manual | | Networking | 1: `dns-dig-action` | No service needed | | Nosql | 2: `couchbase-sink`, `mongodb-changes-stream-source` | Couchbase container, MongoDB as a replica set (for change streams) | | PQC | 2: `pqc-kem-action`, `pqc-signature-action` | No service needed | | Salesforce | 3: `salesforce-composite-upsert-sink`, `salesforce-pubsub-source`, `salesforce-source` | An HTTP mock where the Kamelet takes a base URL; otherwise manual | | Search | 4: `elasticsearch-index-sink`, `opensearch-index-sink`, `solr-sink`, `solr-source` | Elasticsearch, OpenSearch and Solr containers | | Social | 2: `telegram-sink`, `telegram-source` | An HTTP mock where the Kamelet takes a base URL; otherwise manual | | System | 1: `exec-sink` | No service needed | | Transformation | 9: `freemarker-template-action`, `jolt-transformation-action`, `jslt-action`, `mustache-template-action`, `mvel-template-action`, `pdf-action`, `string-template-action`, `velocity-template-action`, `xj-template-action` | No service needed | ### How to go about it - **With #3117:** each batch of #3117 (data types, one family per PR) runs the Kamelets of its family before and after anyway. So it adds the Citrus tests of that family in the same PR. - **The rest:** one PR per family, as for the documentation (#3141 to #3155). - **Services that cannot run in CI** (Snowflake, Databricks, Bedrock, the Google Workspace APIs, Watson): a manual test against the real service, written down in the PR. No `verified` label for these. _Claude Code on behalf of Claus Ibsen_ -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
