* fix: not ready metrics for some edge case conditions on ces
Signed-off-by: Gustavo Carvalho <gusfcarvalho@gmail.com>
* fix: failure conditions with no metrics
Signed-off-by: Gustavo Carvalho <gusfcarvalho@gmail.com>
---------
Signed-off-by: Gustavo Carvalho <gusfcarvalho@gmail.com>
Co-authored-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* feat: implement a cluster-wide generator
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* remove unneeded function
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* check diff run output
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* alternative implementation of the Generator approach using specs only
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* refactor the extracting code
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* slight modification to the naming of the spec from generatorSpec to simply generator
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* write a unit test for the generator and register it in the scheme
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* add documentation for the cluster generator
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
---------
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* feat: enable concurrent reconciling for push secret reconciler
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* add cluster secret store concurrent option as well
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
---------
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
This removes the need for an intermediary Kind=ExternalSecret and
Kind=Secret when using a generator.
Signed-off-by: Moritz Johner <beller.moritz@googlemail.com>
* chore: update go version of the project to 1.23
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* fixed an absurd amount of linter issues
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
---------
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* feat: increase verbosity of error message during validation
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* removing Equal as we do not have the specific error message there
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
---------
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* fix: explicitly fetch status subresource due to inconsistencies
Signed-off-by: Moritz Johner <beller.moritz@googlemail.com>
* fix: bump go
Signed-off-by: Moritz Johner <beller.moritz@googlemail.com>
* fix: add rbac to get status
Signed-off-by: Moritz Johner <beller.moritz@googlemail.com>
---------
Signed-off-by: Moritz Johner <beller.moritz@googlemail.com>
* feat: Add component labels to custom resource definitions
Prerequisite for restricting the CRDs cached by Informer
Signed-off-by: Tsubasa Nagasawa <toversus2357@gmail.com>
* feat(certcontroller): Allow restricting CRDs and Webhook configs in Informer cache
The certcontroller watches CRDs and Webhook configurations, and
manages CA certificates for conversion webhooks of CRDs and Webhook
configurations. Some clusters have a large number of CRDs and Webhook
configurations installed. Additionally, some CRDs have large object sizes.
Currently, the certcontroller holds all CRDs and Webhook configurations
in the Informer cache. Since this includes CRDs not managed by the
certcontroller for CA certificates, memory usage tends to be high.
This PR adds a label to the CRDs and configures the Informer cache to hold
only the CRDs and Webhook configurations restricted by the label selector.
It assumes that the CRDs have a label. Depending on how the External Secrets
Operator is managed, it may be possible to update the External Secrets
Operator without updating the CRDs, so as a precaution, it can be turned
on/off via a startup option. It is disabled by default.
Signed-off-by: Tsubasa Nagasawa <toversus2357@gmail.com>
---------
Signed-off-by: Tsubasa Nagasawa <toversus2357@gmail.com>
* Start reconciliation when a secret has changed
Signed-off-by: shuheiktgw <s-kitagawa@mercari.com>
* Prolong the test timeout
Signed-off-by: shuheiktgw <s-kitagawa@mercari.com>
* Use predicate.ResourceVersionChangedPredicate instead
Signed-off-by: shuheiktgw <s-kitagawa@mercari.com>
---------
Signed-off-by: shuheiktgw <s-kitagawa@mercari.com>
* Address !3331 and !3080
* Modify webhook provider TLS config to restrict tls renegotiation to once per client
** Addresses !3331
* Modify webhook certs validation to include intermediates held within tls.crt
** Addresses !3080
** [Cert-Manager recommendation](https://cert-manager.io/docs/configuration/ca) for CA issuer
Signed-off-by: Rick Mulder <rickymulder@gmail.com>
* Add tls chain tests related to #3080
Signed-off-by: Rick Mulder <rickymulder@gmail.com>
* Clean up tls chain test based on sonarcloud recommendation
Signed-off-by: Rick Mulder <rickymulder@gmail.com>
* Fix checkEndpoints and checkCRDs to use proper pointer reference
Signed-off-by: Rick Mulder <rickymulder@gmail.com>
---------
Signed-off-by: Rick Mulder <rickymulder@gmail.com>
* feat: add push secret to e2e tests
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* finally, a fully working example for an e2e flow with push secret
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* fix value field duplication issue
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
---------
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* feat: add templating to PushSecret
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* adding unit tests around templating basic concepts and verifying output
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* extracting some of the common functions of the parser
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* remove some more duplication
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* removed commented out code segment
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* added documentation for templating feature
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* simplified the templating for annotations and labels
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
---------
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* feat: allow pushing the whole secret to the provider
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* add documentation about pushing a whole secret
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* disabling this feature for the rest of the providers for now
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
* added scenario for update with existing property
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
---------
Signed-off-by: Gergely Brautigam <182850+Skarlso@users.noreply.github.com>
readiness probes are being executed independently from the
leader election status. The current implementation depends on
leader election (client cache etc.) to run properly.
This commit fixes that by short-circuiting the readiness probes
when the mgr is not the leader.
This bug surfaces when `leader-election=true` and cert-controller `replicas>=2`.
Signed-off-by: Moritz Johner <beller.moritz@googlemail.com>
fix: deprecate sourceRef.generatorRef from .data[]
A generator is supposed to be used via .dataFrom[]. Usage in .data[]
is not implemented and doesn't make sense, see #2720.
This commit splits the SourceRef into two types:
- one that only defines a secretStoreRef
- one that allows to define either secretStoreRef or generatorRef
The former is used in .data[] and the latter is used in .dataFrom[].
The Deprecated field is going to be removed with v1.
Signed-off-by: Moritz Johner <beller.moritz@googlemail.com>
* Report not ready when no namespace matches
Signed-off-by: shuheiktgw <s-kitagawa@mercari.com>
* Fix flaky a test
Signed-off-by: shuheiktgw <s-kitagawa@mercari.com>
* Simplify ClusterExternalSecret status
Signed-off-by: shuheiktgw <s-kitagawa@mercari.com>
---------
Signed-off-by: shuheiktgw <s-kitagawa@mercari.com>
