fix: hide 5xx error message from client

[liangliang4ward]

Description

hidding 5xx error message from client

Fixes #6699

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)

[leslie-tsang]

Hello there, I will take over this PR as it lacks of feedback. :)

[liangliang4ward]

Hello there, I will take over this PR as it lacks of feedback. :)

I'm sorry about that, I'm already wait for reply.

now I see you commit , I learn something.

[leslie-tsang]

I'm sorry about that, I'm already wait for reply.

now I see you commit , I learn something.

Ohhh, I see, Glad to have you back, do you like to continue ?

[liangliang4ward]

I'm sorry about that, I'm already wait for reply.
now I see you commit , I learn something.

Ohhh, I see, Glad to have you back, do you like to continue ?

let me do this continue;
I thought that recording ERROR logs might result in two error logs for the test case. Because it's already been recorded once.

Now that I see your commit, I think logging WARN would be a better way.

[leslie-tsang]

let me do this continue; I thought that recording ERROR logs might result in two error logs for the test case. Because it's already been recorded once.

Now that I see your commit, I think logging WARN would be a better way.

Ok, please check those review comments carefully and fix them according to the comments. :)

Would you take some time to finish it today ?

[liangliang4ward]

let me do this continue; I thought that recording ERROR logs might result in two error logs for the test case. Because it's already been recorded once.
Now that I see your commit, I think logging WARN would be a better way.

Ok, please check those review comments carefully and fix them according to the comments. :)

Would you take some time to finish it today ?

ok

[leslie-tsang]

Hello there, @liangliang4ward, There are still some comment need to be take care of. I will reopen them for you.

[liangliang4ward]

@leslie-tsang hi, ci run failed. but not relate this PR code;

[leslie-tsang]

@leslie-tsang hi, ci run failed. but not relate this PR code;

I will check it out. Thanks for you contribution.

[leslie-tsang]

@leslie-tsang hi, ci run failed. but not relate this PR code;

I will check it out. Thanks for you contribution.

CI pass now.

I think we also need to check the error log?

Please add the test case to cover the change.

[liangliang4ward]

hi, I will close this pr, I find the master already modify some code relate with "authz-keycloak.lua" ,some logic is changed.

[tzssangglass]

hi @liangliang4ward , if what you think already modify some code relate with "authz-keycloak.lua" is this PR: #6854, I think there is some misunderstanding.

#6854 does not conflict with this PR, so you can go on and reopen this PR.

[liangliang4ward]

some code I don't know how process;
like function "authz_keycloak_get_token_endpoint"
the master new code call this function return :

return 503, {message = err}  

and other code return:

 return 500, "Unable to determine token endpoint."

and refer from #6699
will be: return 503;

[tzssangglass]

some code I don't know how process; like function "authz_keycloak_get_token_endpoint" the master new code call this function return :

return 503, {message = err}  

and other code return:

 return 500, "Unable to determine token endpoint."

and refer from #6699 will be: return 503;

we can just use 503? cc @spacewander

[spacewander]

It's recommended to return 503 as people keep thinking the 500 status code is the error of APISIX.