Currently the install_trust option for CAs from the pki module can only be configured via JSON/API.

Setting TLS to internal with option on_demand is easily possible with a Caddyfile and allows for a quick and simple internal setup. Trust store installation on the machine itself is not necessary and only leads to startup errors being logged from both sudo and caddy.

Therefore it w

Next line returns 200 as status https://github.com/Kong/kong/blob/c817fada7a514d84f6e0aab7461605c71bf3a841/kong/plugins/cors/handler.lua#L13

Internet says it should return 204:

• https://developer.mozilla.org/en-US/docs/Glossary/Preflight_request
• https://docs.microsoft.com/en-us/aspnet/core/security/cors?view=aspnetcore-3.1#preflight-requests

Internet changed its mind since 2018 https://

On contexts with many APIs and APIs calling other APIs in chains, having too many spans in the zipkin/jaeger dashboards may lead to a too complex usage of such dashboards and of not necessary big data volumes.

For Tyk's admin team, the steps representing the time spent in each Tyk middleware are much valuable.

But the upstream APIs teams just want to know the global time spent in Tyk interna

Improve Docs

https://github.com/apache/apisix/blob/master/FAQ.md#why-is-it-that-installing-apisix-dependencies-with-luarocks-causes-timeout-slow-or-unsuccessful-installation

luarocks install apisix --server https://luarocks.cn

Please describe which part of docs should be improved or typo fixed

A clear and concise description of what you want and what your use case is.

Background

• BFE can be configured with Passive Health Check (i.e. Outlier Detection) and use responses from real requests to determine whether an backend instance is healthy.

Description
Add Active Health Check support:

• Actively and periodically send health checking requests to backend.
• Use responses from health checking requests to determine whether an backend instance is he

Is your feature request related to a problem? Please describe.

As observed during #261, users can accidentally set timeouts inconsistently. A concrete example is that one can set a User facing timeout lower than any backend timeout, such that requests may fail early.

Describe the solution you'd like

We should provide some sanity checking to prevent unintended timeout behavior.

Currently we have a mix of:

• Out of memory error.
• Out of memory error
• out of memory error
• out of memory
• out of memory.
• out of memory [some context]

All these should be adjusted to use the same wording for consistency and it should also make the executable a little bit smaller 😄