S3 sink connector exposes unknown parameter kafka.max.partition.validation.disable

[jonathanschnabelwhatnot]

I have the following S3 sink connector configuration:

resource "confluent_connector" "myosin" {
  environment {
    id = module.kafka.environment.id
  }
  kafka_cluster {
    id = module.kafka.cluster.id
  }

  config_sensitive = {
    "aws.access.key.id"     = aws_iam_access_key.confluent.id
    "aws.secret.access.key" = aws_iam_access_key.confluent.secret
  }

  config_nonsensitive = {
    topics                                   = "mytopic"
    "input.data.format"                      = "JSON"
    "connector.class"                        = "S3_SINK"
    name                                     = "my connector"
    "kafka.auth.mode"                        = "SERVICE_ACCOUNT"
    "kafka.service.account.id"               = ...
    "s3.bucket.name"                         = ...
    "s3.ssea.name"                           = "aws:kms"
    "s3.sse.kms.key.id"                      = ...
    "s3.part.size"                           = "5242880"
    "s3.path.style.access.enabled"           = "true"
    "s3.region"                              = local.region
    "output.data.format"                     = "JSON"
    "json.decimal.format"                    = "NUMERIC"
    "path.format"                            = "YYYY/MM/dd/HH"
    "time.interval"                          = "HOURLY"
    "rotate.schedule.interval.ms"            = "-1"
    "flush.size"                             = "1000"
    "behavior.on.null.values"                = "ignore"
    "timezone"                               = "UTC"
    "subject.name.strategy"                  = "TopicNameStrategy"
    "tombstone.encoded.partition"            = "tombstone"
    "locale"                                 = "en"
    "s3.schema.partition.affix.type"         = "NONE"
    "schema.compatibility"                   = "NONE"
    "store.kafka.keys"                       = "true"
    "value.converter.connect.meta.data"      = "true"
    "store.kafka.headers"                    = "false"
    "s3.object.tagging"                      = "false"
    "value.subject.name.strategy"            = "TopicNameStrategy"
    "max.poll.interval.ms"                   = "60000"
    "max.poll.records"                       = "500"
    "tasks.max"                              = 1
  }
}

The connector deploys correctly, but when I run a subsequent Terraform plan without changing the connector code, it shows changes on the parameter kafka.max.partition.validation.disable.

This parameter is not documented and it seems it should not be exposed to users.

Terraform version: 1.8 Confluent provider version: 1.47

[channingdong]

This issue is ack'ed and fix should be contained in the next official release.

[linouk23]

@jonathanschnabelwhatnot, we deployed a client-side fix for this issue in our latest release, 1.77.0, of the TF Provider 🫡.