Skip to content
Draft
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
132 changes: 132 additions & 0 deletions docs/user-manual/modules/ROOT/pages/camel-4x-upgrade-guide-4_24.adoc
Original file line number Diff line number Diff line change
Expand Up @@ -9970,3 +9970,135 @@ an empty value counts as unset.
The `region` option is no longer marked as required in the catalog, since `region` or `endpoint` is required and the
`kmsClient` option replaces both. The producer still fails with "Region or endpoint is required" when it has to create
a client and neither is set.

=== Kamelets - removed Kamelets

The Kamelet catalog (camel-kamelets 4.24) removes Kamelets that could not work on Camel 4.23 or 4.24:

[cols="2,3",options="header"]
|===
| Removed Kamelet | Use instead
| `twitter-directmessage-source`, `twitter-search-source`, `twitter-timeline-source` | Nothing: the Twitter/X v1.1 API they call is gone or paid-only
| `chunk-template-action` | `mustache-template-action`, `velocity-template-action`, `freemarker-template-action` or `string-template-action`
| `djl-image-to-text-action` | Nothing
| `aws-sts-assume-role-action` | Nothing
| `kafka-azure-schema-registry-sink`, `kafka-azure-schema-registry-source`, `kafka-batch-azure-schema-registry-source` | The schema registry options of the Kafka Kamelets
| `dropbox-source`, `dropbox-sink` | The `dropbox` component
|===

It also removes Kamelets that were little used and did not work well. Use the Camel component instead:

* `kubernetes-pods-source`, `kubernetes-nodes-source`, `kubernetes-namespaces-source`
* `azure-storage-blob-changefeed-source`
* `elasticsearch-search-source`, `opensearch-search-source`
* `cassandra-source`
* `sap-netweaver-sink`
* `google-functions-sink`
* `caffeine-action`

It removes the variants whose options the plain Kamelets now have:

[cols="2,2,3",options="header"]
|===
| Removed Kamelet | Use instead | What to change
| `http-secured-sink`, `http-secured-source` | `http-sink`, `http-source` | Only the name. The login options are the same
| `jms-pooled-apache-artemis-sink` | `jms-apache-artemis-sink` | Only the name. The plain sink now pools its connections, with the same pool options
| `jms-pooled-apache-artemis-source` | `jms-apache-artemis-source` | The name, and remove the three pool options
|===

The demo Kamelets `bitcoin-source`, `chuck-norris-source`, `earthquake-source` and `wttrin-source` are no longer in the
catalog. They are kept as examples in the `examples/kamelets` directory of the camel-kamelets repository. To run one, use
`camel run route.yaml --local-kamelet-dir=examples/kamelets`.

The public `KameletPrefixSchemeEnum` of `camel-kamelets-catalog` loses the entries of these Kamelets: `dropbox`, the
`twitter_*` entries, `azure_storage_blob_changefeed`, `elasticsearch_search`, `google_functions`, the `kubernetes_*`
entries, `bitcoin`, `chuck_norris`, `wttrin` and `http_secured`.

=== Kamelets - deprecated Kamelets

These Kamelets still work, and will be removed in a later release:

[cols="2,2,3",options="header"]
|===
| Deprecated Kamelet | Use instead | What to change
| `kafka-apicurio-registry-not-secured-sink`, `kafka-apicurio-registry-not-secured-source`,
`kafka-batch-apicurio-registry-not-secured-source`, `kafka-batch-apicurio-registry-source`,
`kafka-not-secured-apicurio-registry-json-source`, `kafka-not-secured-apicurio-registry-sink`,
`kafka-not-secured-apicurio-registry-source`
| `kafka-source`, `kafka-sink` and `kafka-batch-source` | Use the schema registry options of the Kafka Kamelets
|===

=== Kamelets - renamed options

These Kamelet options had the name of a parameter that the `kamelet:` component takes for itself. Given on a `kamelet:`
endpoint, they never reached the Kamelet: `location` was read as the place to load the Kamelet from, so the route failed
to start. Set as `camel.kamelet.<name>.<option>` properties, such as in a Camel K Pipe, they worked, so rename them:

[cols="2,1,1",options="header"]
|===
| Kamelet | Old name | New name
| `google-vertexai-sink` | `location` | `region` (required)
| `ibm-cos-sink`, `ibm-cos-source` | `location` | `region`
| `fhir-sink`, `pulsar-sink` | `lazyStartProducer` | `lazyStart`
|===

With the old name, `google-vertexai-sink` fails at startup because `region` is missing, and the other Kamelets ignore it.

=== Kamelets - Kafka Kamelets require saslAuthType

`kafka-sink`, `kafka-batch-source` and `ceph-event-based-source` have no default for `saslAuthType` any more, as was
already the case for `kafka-source`. Without the option, the route fails at startup:

[source,text]
----
Route template kafka-sink the following mandatory parameters must be provided: saslAuthType
----

To connect without security, as before, set `saslAuthType: NONE`.

=== Kamelets - messaging sinks no longer wait for a reply

`jms-apache-artemis-sink`, `jms-amqp-10-sink`, `jms-ibm-mq-sink`, `spring-rabbitmq-sink` and `nats-sink` now send a message without waiting for a reply, also when the message expects
one, such as a message from `webhook-source`. The caller gets its own message back, instead of the answer of a
receiver.

Other changes to the messaging Kamelets:

* `jms-apache-artemis-sink` pools its connections: it opens one connection to the broker instead of one per message.
`jms-apache-artemis-sink` and `jms-apache-artemis-source` have new `username` and `password` options.
* The body of `nats-source` is a `byte[]` instead of a stream. The `Content-Type` of the sender is kept instead of
being set to `application/json`.
* `mqtt-source` and `mqtt5-source` without a `clientId` connect with a generated client ID instead of `mqtt-source`.
This matters only for broker ACLs or monitoring that use that ID.
* `spring-rabbitmq-source` and `spring-rabbitmq-sink` with `protocol: amqps` check the certificate of the broker
against the trust store of Java, and that it names the host in `host`. A broker whose certificate Java does not
trust, or that is for another name, no longer connects. Add its CA to the trust store of Java, or give it with the new
`sslTruststoreLocation` and `sslTruststorePassword` options. `protocol: amqp`, the default, is unchanged.
* `pulsar-sink` also takes the message key from a plain `key` header, which `pulsar-source` sets, so a key carries
over from one Kamelet to the other. When both are set, the `CamelPulsarProducerMessageKey` header wins. Messages that
had a `key` header for another purpose now use it as the Pulsar key.

=== Kamelets - other changes in behaviour

* `solr-source` gives JSON text, with `Content-Type: application/json`, instead of a SolrJ object. A route that read
the fields of the object must read the JSON instead, for example with `jsonpath`.
* The SQL and Cassandra Kamelets send a query with `+` to the database unchanged. A query that worked around this, for
example with `%2B`, now sends a literal `%2B`.
* `google-sheets-source` with `majorDimension=COLUMNS` reads columns.
* `mongodb-changes-stream-source` sends the whole document for an update, read again from the collection, instead of
`null`. Set the new `fullDocument` option to `default` for the old behaviour.
* `redis-sink` sends a PUBLISH that has no channel header to the channels of the `channels` option (default `one`),
instead of failing.
* The GitHub source Kamelets poll every 60 seconds instead of every 500 milliseconds. Set `delay` to poll more often.
* The Jira sink Kamelets give an issue the description of the `ce-issueDescription` header when the body is empty.
* `rest-openapi-sink` with `clientRequestValidation=true` rejects an invalid request instead of sending it.
* In `mail-sink`, a `ce-to` or `ce-cc` header replaces only its own option. Before, either of them dropped all the
recipient options, including bcc.
* `aws-translate-action` no longer requires `sourceLanguage`. With neither it nor `autodetectSourceLanguage`, the
Kamelet starts, and each message fails.
* `scp-sink` takes a path or a `classpath:` resource in `knownHostsUri`. A `file://` or `http(s)://` value, which
never worked, is rejected.
* `header-matches-filter-action` requires `headerName`, `jira-source` requires `jql`, `salesforce-create-sink`
requires `sObjectName` and `fhir-sink` requires `apiName`. These Kamelets could not start without them, and now
fail with a clear error. `aws-bedrock-agent-sink` requires `dataSourceId`, and fails at startup without it instead
of on every message.
Loading