feat(plugin)(correlationId) increase priority order for plugin to mat… #13581
Conversation
…ch EE so it can be used in other/custom plugins
github-actions
bot
added
plugins/correlation-id
cherry-pick kong-ee
team-eng-enablement
added
the
author/community
|
I patch my existing correlation id plugin in our own runtimes to execute earliest too as it makes sense to be able to capture that in other http logging plugins etc. |
…value is null KAG-5287
…eir values mismatch (Kong#13565) * fix(*): reject config if both deprecated and new field defined and their values mismatch This commit adds a validation to deprecated fields that checks if in case both new field and old field were defined - their values must match. Note that if one of the fields is null then the validation passes even if the other is not null. KAG-5262 * fixup! fix(*): reject config if both deprecated and new field defined and their values mismatch PR fixes * fixup! fix(*): reject config if both deprecated and new field defined and their values mismatch PR fixes 2
add an action to enable executing the `/prdiff <other-pr-url>` command, to get the diff between the current PR's changes and the other PR's. This gives a quick overview of the differences between two PRs and is meant to facilitate the process of reviewing cherry-picks.
### Summary Adds libada and lua-resty-ada as a dependency. This is needed for: Kong#12758 But it may be great for many other uses too. The `lua-resty-ada` LuaJIT FFI bindings can be found here: https://github.com/bungle/lua-resty-ada Signed-off-by: Aapo Talvensaari <[email protected]>
…t_redis_connection()` are mistaken (Kong#13613) * fix(rate-limiting-plugin): fix a bug where the return values from `get_redis_connnection()` are mistaken * fix changelog
It is one of the serial refactors of helpers.lua KAG-5236 The function `kong_exec()` needs an upvalue `conf`, which must be reloaded each time.
It is one of the serial refactors of helpers.lua This function (reload_module) may be used by test cases outside.
It is one of the serial refactors of helpers.lua KAG-5242
…ong#13610) This PR modifies the `cache_key` function of the vault entity to always generate a cache key without workspace id. Vault entity is workspace-able, but our secret rotation timer always run without workspace settings(thus the default workspace is being used), so during secret rotation, the code https://github.com/Kong/kong/blob/4e38b965b922f57febe8652fb96b7d74aeab591a/kong/pdk/vault.lua#L620-L621 will generate a duplicate vault cache with default workspace id for each non-default workspace vault entity, and those cache will never be refreshed. The result of this issue is that when you update a vault entity's configuration inside a non-default workspace, it will never take effect in the secret rotation. Since the prefix of vault entity is unique across workspaces, it should be safe to only use one cache key without workspace id, so that the correct cache is used during secret rotation. FTI-6152
A small PR for adding a null handling code for the multiValueHeaders field. Also contains a small refactor to put isBase64Encode check into the validate function to keep consistency with other fields FTI-6168
…lugins iterator (Kong#13602) the workspace id is needed in some cases, for example in the configure handler. https://konghq.atlassian.net/browse/FTI-6200
github-actions
bot
added
plugins/aws-lambda
plugins/acme
core/db
plugins/response-ratelimiting
chore
|
Can we do this in |
Generally, changing plugins ordering is a breaking change and should be avoided in a non-major; especially if the change decreases the priority - ie, its side effects cease being visible to plugins it previously was visible. In this particular case, this does not happen: it expands the list of plugins that have access to the id. Also, there's no change in functionality. I would say it's okay to merge in a minor or even a patch. |
|
@gszr - when do you think this will make it in a version of CE? |
@andrewgkew This will likely be included in 3.9, which is planned for late December. |
|
I think @bungle made a good point, technically this is a breaking change, but if the EE version already increased the priority and we have not heard much complaints, it should be enough proof that this user probably doesn't notice the difference. Wouldn't consistency be more of a consideration here? |
|
@andrewgkew could you add a changelog entry under |
|
Cherry-pick failed for Please cherry-pick the changes locally. git remote add upstream https://github.com/kong/kong-ee
git fetch upstream master
git worktree add -d .worktree/cherry-pick-13581-to-master-to-upstream upstream/master
cd .worktree/cherry-pick-13581-to-master-to-upstream
git checkout -b cherry-pick-13581-to-master-to-upstream
ancref=$(git merge-base 7b4f198a8b322d0d6c33767add195bfcb275f743 c6896701248d10f5b53631dacf4a88c44ab976e5)
git cherry-pick -x $ancref..c6896701248d10f5b53631dacf4a88c44ab976e5 |
github-actions
bot
added
the
incomplete-cherry-pick
kikito
removed
cherry-pick kong-ee
|
Since this change already exists in kong ee, cherrypicking is not necessary. Removing the cherrypick labels. |
Summary
Increase priority order for plugin to match EE so it can be used in other/custom plugins
Checklist
changelog/unreleased/kongorskip-changeloglabel added on PR if changelog is unnecessary. README.mdIssue reference
Fix #[issue number]