Telegraf not writing to InfluxDB if buffer for CrateDB fills up to metric_batch_size about telegraf HOT 2 CLOSED

If the outputs.cratedb buffer fills up to metric_batch_size, the oldest metrics are expected to be dropped. Regardless of this, InfluxDB should never stop writing

Unfortunately, this is not possible when using a service input like MQTT Consumer. Service input plugins use tracking metrics that ensure the delivery of metrics from the input all the way to the output and notify the original source once delivered.

Additionally, these types of inputs will only read up to max_undelivered_messages, which is 1000 by default. This is because without some sort of limit you could easily overwhelm telegraf and whatever outputs you are using by collecting and sending a huge number of messages all at once.

If your instance has read up to 1000 metrics, which is what your buffer reports, and the messages are unable to be delivered, the mqtt_consumer plugin will not continue to read any further messages until that changes.

One possible alternative is to have two separate mqtt plugins one for each output.

from telegraf.

Ok, so the problem is with the mqtt consumer. Makes sense. Using two inputs could be a valid workaround for me.
Thanks!

from telegraf.