fix: fix consul plugin worker will not exit while reload or quit

[theweakgod]

Description

1.configure

config.yaml

discovery:
consul:
servers:
- "http://localhost:8500"
keepalive: true

Version: 3.6.0

2.situation
Using the consul service, I found that there is a process leak. When reloading or quitting, there will be a nginx process shutting down and never exit.
When nginx reloads, the consul worker with keepalive turned on will not exit normally, will be in shutting down state, and will always access consul's health api.
You only need to make a judgment before creating a new timer to solve the process leak.
This problem does not exist when keepalive is false, but there will be an obvious memory leak.

htop -p $(pgrep -d',' -f "nginx")

process will not exit!

To process exit when testing, please change consul node state.

3.testing
start consul discovery, and try to apisix reload && htop -p $(pgrep -d',' -f "nginx"). The process in shutting down, and nerver exit even if you set timeout.
after pull this commit, set shutting down timeout 300 seconds or change consul node info, the shutting down process will exit after node info changing or timeout.

Fixes # (issue)

Checklist

• I have explained the need for this PR and the problem it solves
• I have explained the changes or the new features added to this PR
• I have added tests corresponding to this change
• I have updated the documentation to reflect this change
• I have verified that this change is backward compatible (If not, please discuss on the APISIX mailing list first)

[theweakgod]

During testing, if nginx worker is still shutting down after reload, you can update the consul node so that the worker will exit, because the worker is in the http request state at this time and will exit after the request is completed.😬

[shreemaan-abhishek]

I checked this, the worker process still does not exit. Do you have a clue on how we can make the worker process exit automatically instead of worker reset?

[theweakgod]

I checked this, the worker process still does not exit. Do you have a clue on how we can make the worker process exit automatically instead of worker reset?

You can update consul node(eg. register a server). Because worker is blocked by http request for monitoring consul node update. Then worker will exit.
Or waiting a few minutes. worker will exit also.
reference

[theweakgod]

During testing, if nginx is still shutting down after reload, you can update the consul node so that the worker will exit, because the worker is in the http request state at this time and will exit after the request is completed.😬

Check this. @shreemaan-abhishek