add: --ready flag for postgrest healthcheck

[taimoorzaeem]

The --ready flag is a wrapper around the admin server /ready request. This is done through using an http client library in postgrest.

• Implementation
• Docs

Closes #4239.

[taimoorzaeem]

@steve-chavez Where do we document this new flag, I guess in https://docs.postgrest.org/en/v13/references/cli.html?

[steve-chavez]

@taimoorzaeem Yes, should be in CLI.

[steve-chavez]

@taimoorzaeem Just tried:

$ postgrest-run -- --ready
postgrest: HttpExceptionRequest Request {
  host                 = "0.0.0.0"
  port                 = 3001
  secure               = False
  requestHeaders       = []
  path                 = "/ready"
  queryString          = ""
  method               = "GET"
  proxy                = Nothing
  rawBody              = False
  redirectCount        = 10
  responseTimeout      = ResponseTimeoutDefault
  requestVersion       = HTTP/1.1
  proxySecureMode      = ProxySecureWithConnect
}
 (ConnectionFailure Network.Socket.connect: <socket: 13>: does not exist (Connection refused))

The error is expected, but can we make it clearer? (no Haskell type wrapping)

[steve-chavez]

Also noted that on success we don't print anything:

$ postgrest-run -- --ready

# result is good
$ echo $?
0

Maybe we should print something like traefik:

$ postgrest-run -- --ready
OK: http://:<port>/ready

[taimoorzaeem]

This test is flaky and fails sometimes because the subprocess.run exits without capturing the stderr output. I guess this happens due to output not being flushed properly. Any ideas on how to handle this cleanly?

[steve-chavez]

Maybe instead try to force a schema cache failure so /ready always fails. An idea could be to put a bad database connection uri. Check the io tests too, likely this was done before.