Thanos querier http_requests_total metric reports 503 status code

[yaron-idan]

Hey,
We are using Thanos to aggregate results from multiple k8s clusters.
Lately we've been seeing a rise in 503 error status codes emitted from the query container in our thanos-external-querier pod . However, when we look at the logs - there doesn't seem to be a log correlated to these errors.
We've tried to increase verbosity of these components but even with debug logs there is no indication of the request causing these errors.
What would be the proper way to troubleshoot this issue?