Expected timestamp or new record, got MNAME


I’m trying to debug an error with app exported Prom metrics; getting:

prom-prometheus-server-6979dc7bb-t9jm9 prometheus-server ts=2022-01-04T21:28:20.943Z caller=scrape.go:1307 level=debug component=“scrape manager” scrape_pool=kubernetes-pods target= msg=“Append failed” err=“expected timestamp or new record, got “MNAME””

I can’t visually see what line this is on from reading the same endpoint that is failing.

Doing a

curl --silent localhost:15020/stats/prometheus | promtool check metrics

Yields no extra information either besides naming suffix suggestions.

How can I find 1) what line is causing this and what its contents are?, 2) how to fix it?

(I’m using the vanilla prometheus_client exporter, with three simple counters and one Gauge, as well as Istio’s merged metrics.)