Cloud Events Server
How to run
With sqlite as backend database:
go run --tags sqlite3 . -config=configs/example-config-sqlite3.yaml
With MySQL or TiDB as backend database:
go run . -config=configs/example-config.yaml
Tekton step-log fetching was removed
Tekton failure notifications no longer read pod logs. The fetched target used to
come from the CloudEvent payload (namespace, podName, containerName), and
POST /events is unauthenticated, so any producer could turn the notifier into
an arbitrary pod-log reader and have the text delivered to a Lark recipient of
its choosing. The capability was also already dead in production: the Tekton
Dashboard sits behind Envoy Gateway's Google OIDC policy, and the fetch carried
no credentials and refused redirects, so it received a 302 and treated it as an
error.
Cards keep the failed step names and Terminated.Reason, the dashboard view
link built from tekton.dashboard_base_url, and the rerun command; read the
logs on the Dashboard. The tekton.failed_step_tail_lines setting is gone from
the config schema.
How to release
- manualy build the container images with docker
skaffold build --default-repo <your image registry>
- manualy build the container images with cluster(that has amd64 and arm64 nodes)
skaffold build --default-repo <your image registry>
more...