Gimbal Versions Save

Gimbal is an ingress load balancing platform capable of routing traffic to multiple Kubernetes and OpenStack clusters. Built by Heptio in partnership with Actapio.

v0.4.0

5 years ago

Discoverer image: gcr.io/heptio-images/gimbal-discoverer:v0.4.0

Merged pull requests:

  • openstack-project-watchlist: List of projects to be watched for reconcilation. If empty, load balancers across all projects will be reconciled #249 (@hmorikaw)

v0.3.1

5 years ago

Change Log

Discoverer image: gcr.io/heptio-images/gimbal-discoverer:v0.3.1

Merged pull requests:

  • pkg/openstack: enqueue the actual svcs instead of a ptr to the range var #231 (alexbrand)
  • Replace special char in openstack lb names when necessary and remove lb name validation #232 (alexbrand)

v0.3.0

5 years ago

Change Log

Discoverer image: gcr.io/heptio-images/gimbal-discoverer:v0.3.0

New and Improved

Openstack Discovered Name

The openstack translation process now uses the load balancer's ID as the service/endpoint name. The load balancer name is still available as a label on the service.

Probes

The openstack discoverer now has a Liveliness probe which will restart the discoverer if connection to the Openstack API fails.

Metrics & Dashboards

The metrics from Contour have been updated and the included dashboards have been improved.

Merged pull requests:

v0.3.0-beta.1

5 years ago

v0.3.0_alpha.1

5 years ago

v0.2.0

5 years ago

Discoverer image: gcr.io/heptio-images/gimbal-discoverer:v0.2.0

New and Improved

Advanded load balancing and routing features with the IngressRoute CRD

The initial implementation of the IngressRoute Custom Resource Definition (CRD) has been merged into Contour. Using this resource, you can define ingress rules that load balance requests to a specific Virtual Host across multiple backend clusters. See the design document for more information.

Discovered Service/Endpoints naming scheme changed

The naming scheme used when creating discovered Service and Endpoint resources in the Gimbal cluster has been changed. Once you update the Gimbal discoverers, they will re-discover the existing backend Services and Endpoints using new names. See the discovery naming conventions document for more information.

The Services and Endpoints discovered by the previous version will remain in the cluster, but they will not be updated once the new version of the discoverer is running.

The following command will list all Services and Endpoints that were discovered using the old naming scheme:

kubectl get svc,endpoints --all-namespaces -l gimbal.heptio.com/cluster

The following command will delete all Services and Endpoints in a specific namespace that were discovered by the old discoverers:

kubectl delete svc -n ${NAMSPACE} -l gimbal.heptio.com/cluster

IMPORTANT: Doing this will break any Ingress resources that are referencing these services.

Breaking changes

  • Discoverer configuration flag change: the --cluster-name flag has been removed and replaced by the --backend-name flag.
  • New naming scheme for discovered Services and Endpoints.

Action Required

  • Update Ingress resources to use the new Service names once the discoverer components are updated. This is required because the Service/Endpoint naming scheme has been updated, and old Service and Endpoints resources will not reflect any changes that occur in backend clusters.

Merged Pull Requests

  • Rename cluster-name to backend-name #120 (stevesloka)
  • edit docs for 0.2 release #119 (Bradamant3)
  • Link to Contour tag v0.6.0-alpha.1 #118 (rosskukulinski)
  • grafana: update prometheus query to use new gimbal label #116 (alexbrand)
  • [Doc] Specify the Secret key for the kubeconfig data #114 (rosskukulinski)
  • Sync with Contour IngressRoute version / Add IngressRoute example #112 (stevesloka)
  • Update stats port for prometheus metrics / Add IngressRoute CRD #111 (stevesloka)
  • Update discoverers to conform to new naming schemes #109 (alexbrand)
  • Stop exposing Envoy admin interface #105 (stevesloka)
  • Docs on stateful Prometheus deployment #103 (stevesloka)
  • docs: add discovery naming conventions #102 (alexbrand)
  • Add Queue Size / Openstack API Latency / Openstack Cycle time metrics #97 (stevesloka)
  • Add github issue template #96 (alexbrand)
  • grafana: use kube-state-metrics to count discovered svcs/endpoints #93 (alexbrand)
  • Enable statsd metrics collection for histograms #91 (stevesloka)
  • Split Contour from Envoy pods and implement Network Policy #89 (stevesloka)
  • Add team docs with sample RBAC permissions #88 (stevesloka)
  • Improve performance by precompiling regex #83 (kpango)
  • discovery/pkg/openstack/httplogger.go: fix error message punctuation #79 (kpango)
  • discovery/pkg/openstack/client.go: fix typo #78 (kpango)
  • discovert/cmd/kubernetes-discoverer: remove constant workingNamespace #76 (kpango)
  • discovery/pkg/openstack: remove unused constant #70 (alexbrand)
  • Return error and delete panic code #69 (kpango)
  • Simplify Metrics Definition #68 (kpango)
  • deployment/readme: fix TOC links #65 (alexbrand)
  • discovery/pkg/sync: use single-quotes in log messages #63 (alexbrand)
  • discovery: make the openstack user domain name configurable #61 (alexbrand)
  • Bump contour version to v0.5.0 #59 (stevesloka)
  • Api latency #58 (stevesloka)
  • Implement Cycle Duration Prometheus metric #55 (stevesloka)
  • discoverer: set clustername when creating syncqueue in openstack discoverer #52 (alexbrand)
  • Add link to slack channel #50 (castrojo)
  • Add queue size prometheus metric #49 (stevesloka)

v0.1.0

6 years ago

The initial release of Gimbal!

Discoverer image: gcr.io/heptio-images/gimbal-discoverer:v0.1.0