A couple of config tweaks to fix some small things I noticed:
When using the long poll /v2/status/wait-for-block-after/{round} endpoint the NGINX proxy upstream timeout can end before the server has responded. The timeout has been increased to ensure that won't happen.
Sometimes a 502 was returned when calling KMD. Setting the port in the docker compose file fixes this. Both indexer and algod endpoints are fine.
Tuned config to ensure connections are kept alive.
A couple of config tweaks to fix some small things I noticed:
/v2/status/wait-for-block-after/{round}
endpoint the NGINX proxy upstream timeout can end before the server has responded. The timeout has been increased to ensure that won't happen.