Issues with splitting metrics sent to StatsD - Telegraf

We receive random failures like the below one on the telegraf log. Many of them are successful from the same code, but some are failing. I see a lot of discussion about this happening when the load is high, or when the number of messages limit is reached. But that’s not the case with our systems. It’s happening across environments.

E! Error: splitting ‘:’, Unable to parse metric: _sc|GetRunsList|0

Telegraf Used is 1.5.2

This could be a bug in the statsd input, but often we get reports like this that are caused by issues with statsd clients. Would it be possible to create a packet capture on the Telegraf host while this happens:

tcpdump -i <interface> udp port 8125 -X > statsd_dump