ci: Merge changes from v2 to release 2.8 branch #5379
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
* bump(librdkafka): from v1.9.2 to v2.3.0 - update needed for automatic re-authentication of kafka connections via SASL (used when connecting to Confluent Cloud via OAUTHBEARER) * fix(librdkafka): Copy certs bundle to path expected by librdkafka v2.3.0 librdkafka uses libcurl for fetching OIDC tokens. However, depending on OS, the location where the certs ca-bundle for validating the tls connection between librdkafka and the auth provider is searched differs. We now copy the ca cert bundle to multiple paths to make sure it is found. While this is a bit hacky, there are limited alternatives we can take because: - until now (with librdkafka 1.9.2) those certs were searched under /etc/ssl/certs/ca-certificates.crt but after the update they are searched under /etc/pki/tls/certs/ca-bundle.crt - libcurl doesn't get passed/doesn't respect the caCert location or .pem file set via librdkafka/kafka configs (see confluentinc/librdkafka#375). - there is no way clear way to influence the libcurl search path via env vars (setting the CURL_CA_BUNDLE env var has no effect), with search paths fully decided at compile-time. The likely chain of reasons this has popped-up when updating librdkafka to v2.3.0: - we're building {model, pipeline}gateway docker images based on debian bullseye image - we're copying the build results into redhat ubi9 container - libcurl was statically built in previous librdkafka, so was using the debian search paths on ubi9 - libcurl ends up being dynamically linked in the current librdkafka (to ubi9 version) or has changed its default ca bundle search path
* send server statuses on controller connect * add error message on server notify err * add logs on server notify memory * review comments
Bumps grafana/grafana from 10.3.1 to 10.3.3. --- updated-dependencies: - dependency-name: grafana/grafana dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Bumps [google.golang.org/grpc](https://github.com/grpc/grpc-go) from 1.61.0 to 1.61.1. - [Release notes](https://github.com/grpc/grpc-go/releases) - [Commits](grpc/grpc-go@v1.61.0...v1.61.1) --- updated-dependencies: - dependency-name: google.golang.org/grpc dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
…up (#5353) * set default OTEL_EXPORTER_OTLP_PROTOCOL in compose setup * go mod tidy operator
Bumps [github.com/go-playground/validator/v10](https://github.com/go-playground/validator) from 10.17.0 to 10.18.0. - [Release notes](https://github.com/go-playground/validator/releases) - [Commits](go-playground/validator@v10.17.0...v10.18.0) --- updated-dependencies: - dependency-name: github.com/go-playground/validator/v10 dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
…5358) We need better error handling and reporting across dataflow, together with appropriately cleaning the kafka streams state once a pipeline reaches an error state. Because we're now continuously retrying the subscription to the scheduler when the subscription terminates, not cleaning the kafka streams state of failed pipelines may lead to the inability to re-create the same pipeline, even after deletion. This commit also introduces a new way of handling errors and the changes in pipeline status. It moves us towards being able to inspect the status of the underlying kafka streams from outside the Pipeline object. The end-goal (not archieved here) is to react to pipelines reaching error states during their operation, rather than just when they are created for the first time (what's currently implemented). However, the changes here do get us to a point where the pipeline creation status, including possible errors/exceptions are reported to the scheduler and can show up either in the seldon CLI or in k8s. Instead of propagating exceptions, the idea is to return and handle common errors in a more similar way to golang. **Fixed issues**: - Fixes #INFRA-755 (internal): Exceptions in pipeline creation/deletion are unrecoverable because of uncleaned Kafka Streams state - Progress on #INFRA-617 (internal): Investigate high priority unhandled errors - Progress on #INFRA-648 (internal): Handle dataflow [...] exceptions and present meaningful error messages
significant updates in this pack: - kafka-streams from 3.4.0 to confluent's 7.6.0-ccs
* fix(dataflow) wait for kafka topic creation Kafka topic creation happens asynchronously. This means that even when the return value from `createTopics(...)` indicates that the topic has been created successfully, the topic can not be immediately subscribed to. Instead of verifying the status of the topic from the `createTopics` return value, here we're repeatedly calling `describeTopics` until all of the topics for the pipeline can be described successfully. This indicates that the topic has been fully created _at least_ on one broker, and can now be subscribed to. **Fixed issues**: - Fixes dataflow component for #INFRA-663 (internal): Pipeline creation goes into ERROR state
* fix(gateway) wait for kafka topic creation Kafka topic creation happens asynchronously. This means that even when the return value from createTopics(...) indicates that the topic has been created successfuly, the topic can not be immediately subscribed to. We retry DescribeTopics until all of the topics for the pipeline can be described successfully. This indicates that the topic has been fully created at least on one broker, and can now be subscribed to. Minor changes: - timeout config now in gateway/constants.go - tidying-up of error reporting Which issue(s) this PR fixes: Fixes gateway component for #INFRA-663 (internal): Pipeline creation goes into ERROR state
sakoush
changed the title
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Changes for release 2.8 rc2