Outputs.influxdb did not complete within its flush interval - means points get lost

Telegraf
1

I’m using a telegraf Gateway to write to InfluxDB (v1), there is a warning occasionally popping up in the log, but I’ve always thought nothing of it…

{...} W! [agent] ["outputs.influxdb"] did not complete within its flush interval

Now, after quite some time turns out this “warning” means I’m losing data…
I’ve “fixed” the issue by adjustingflush_interval and batch_size, still I don’t like the idea of losing data due to something that I’d compare to a timeout
Is there a configuration to put those data back in the buffer and write them in the next interval?

2

@Giovanni_Luisotto please report a bug and specify a way to reproduce the issue. Usually the output should put the metrics back into the buffer if it can’t send it, however as the buffer-size is limited it might overflow and thus loose data…

1 Like