TBS: surface pubsub errors and handle errors gracefully #17117
Description
APM Server version (apm-server version): 9.0, but affects all versions
Description of the problem including expected versus actual behavior:
When _refresh request fails, e.g. returning 400, pubsub only logs the error at Debug level. It also stops pubsub from getting latest remote sampling decisions.
Steps to reproduce:
Mock ES to return 400 to _refresh with
{
"_shards": {
"total": 52,
"successful": 31,
"failed": 21,
"failures": [
{
"shard": 0,
"index": "partial-.ds-traces-apm.sampled-default-2025.05.30-001751",
"status": "BAD_REQUEST",
"reason": {
"type": "illegal_argument_exception",
"reason": "Validation Failed: 1: segment generation is unknown;"
}
},
...
Provide logs (if relevant):
Metadata
Metadata
Assignees
Type
Fields
Give feedbackNo fields configured for issues without a type.