Horizon: improve logging when config is broken

[jacekn]

What problem does your feature solve?

It would make troubleshooting easier

What would you like to see?

I had misconfigured horizon where it was unable to reach history archive defined using the HISTORY_ARCHIVE_URLS variable.
Horizon started but it didn't bring up captivecore.
In the logs I had lots of messages similar to:

time="2024-07-17T11:55:02.620Z" level=error msg="failed to load the stellar-core info" err="http request errored: Get \"http://127.0.0.1:11626/info\": dial tcp 127.0.0.1:11626: connect: connection refused" pid=221079 stack="[main.go:43 client.go:139 app.go:239 app.go:459 asm_amd64.s:1695]"
time="2024-07-17T11:55:03.620Z" level=warning msg="could not load stellar-core info: http request errored: Get \"http://127.0.0.1:11626/info\": dial tcp 127.0.0.1:11626: connect: connection refused" pid=221079

which made it harder to spot root cause which was only logged once:

time="2024-07-17T11:54:59.638Z" level=error msg="Error in ingestion state machine" current_state="resume(latestSuccessfullyProcessedLedger=2658758)" error="error preparing range: error starting prepare range: opening subprocess: error calculating ledger and hash for stellar-core run: error getting latest checkpoint sequence: error getting root HAS: Get \"http://<redacted>/.well-known/stellar-history.json\": dial tcp <redacted>:80: i/o timeout" next_state=start pid=221079 service=ingest

I think it would be worth suppressing the captivecore connection refused messages if horizon is stuck talking to the history archives. If the history archive message was the last one logged I think it would be easier to spot root cause in the logs.

What alternatives are there?