fix(core) disable balancer upstream passive health check

[mengskysama]

Full changelog

upstream passive health check can be disable follow doc https://getkong.org/docs/0.12.x/health-checks-circuit-breakers/

Issues resolved

Fix #3304

[hishamhm]

Hi @mengskysama! Thank you for reaching out. I'm still trying to understand the issue better, because I couldn't reproduce it locally, neither by hand or by writing a regression test (add this to the "#healthchecks" section of spec/02-integration/05-proxy/09-balancer_spec.lua):

      it("does not perform health checks when disabled", function()

        local timeout = 10
        -- server responds, then fails, then responds again
        local server = http_server(timeout, localhost, PORT, { 30, 30, 30 })

        -- disable all healthchecks
        local api_client = helpers.admin_client()
        assert(api_client:send {
          method = "PATCH",
          path = "/upstreams/" .. upstream.name,
          headers = {
            ["Content-Type"] = "application/json",
          },
          body = {
            healthchecks = healthchecks_config {
              active = {
                healthy = {
                  interval = 0,
                  successes = 0,
                },
                unhealthy = {
                  interval = 0,
                  tcp_failures = 0,
                  timeouts = 0,
                  http_failures = 0,
                },
              },
              passive = {
                healthy = {
                  successes = 0,
                },
                unhealthy = {
                  tcp_failures = 0,
                  timeouts = 0,
                  http_failures = 0,
                },
              },
            }
          },
        })
        api_client:close()

        local oks, fails, last_status = client_requests(30)
        assert.same(30, oks)
        assert.same(0, fails)
        assert.same(200, last_status)

        oks, fails, last_status = client_requests(30)
        assert.same(0, oks)
        assert.same(30, fails)
        assert.same(500, last_status)

        oks, fails, last_status = client_requests(30)
        assert.same(30, oks)
        assert.same(0, fails)
        assert.same(200, last_status)

        -- collect server results
        local _, server_oks, server_fails = server:join()
        assert.same(60, server_oks)
        assert.same(30, server_fails)

      end)

This works as expected on 0.12.3 and master. So, I'm afraid you're hitting into something else and the behavior you are observing is a side-effect.

Could you try to write a regression test similar to the above with the sequence of steps necessary to perform to trigger the issue?

[mengskysama]

@hishamhm Thanks your reply. I took a moment to confirm the problem and I found that I misunderstood part of the code logic. Problems may occur within the event of cluster, I will give a reproduce code in issue.