Skip to main content

2024-04-16

API / Control Plane

Added

  • Support for default TTL per cluster for Sandboxes and RouteGroups (docs).

2024-04-05

Operator v0.16.1

Fixed

  • Problem during the computation of workload status in forked workload controller. This fixes issue #50.
  • Enforce stricter limits in the names of the forked Argo Rollout objects to avoid the creation of pods with names exceeding 63 characters.

Dashboard

Added

  • Improved layout of Route Groups and Resources detail pages.

Fixed

  • Issue causing 401 Unauthorized errors in Explore UI.

2024-03-06

Operator v0.16.0

Added

  • Local Sandboxes containing local workloads without associated services are now supported.
  • The tunnel-proxy has a new RPC called GetHostnames used by signadot local connect to serve more Service types and operate independently of cluster IP networks.

Fixed

  • When running with Istio enabled, the tunnel-proxy service now has better support for TCP over its socks5 server because the appProtocol field in the associated Service is now correctly configured for Istio.
  • Sandboxes with many forks were causing many unnecessary reconciles and reconciliation conflicts that could lead to hotlooping. This problem was fixed by adding stable orderings to some custom resource structures.

CLI v0.7.0

Added

  • RouteGroups now support TTL.
  • A new sub-command signadot local proxy is available, allowing one to proxy in-cluster services in a way similar to Kubernetes port-forwarding.
  • A new connection type, ControlPlaneProxy for signadot local connect is available, which allows using local without a kubeconfig - intended primarily for onboarding.
  • When running signadot local connect against a cluster with Signadot Operator ≥ v0.16.0, all Kubernetes service names are now available via DNS on the local machine, instead of only clusterIP typed services. Additionally, in the case of StatefulSets with associated headless services, the StatefulSet pod names are available as a subdomain names of the service name pod-0.name.namespace.svc, pod-1.name.namespace.svc etc.
  • When running signadot local connect against a cluster with Signadot Operator ≥ v0.16.0, local workloads with no associated services are now supported.
  • There is a new field in the local section of the config file, called virtualIPNet documented here.

Changed

  • signadot local connect --wait now accepts an optional string argument indicating none, connect or sandboxes, defaulting to connect. These arguments cause the command to wait for nothing, outbound connection success, or outbound connection success and sandbox re-connection success respectively. By default, signadot local connect now behaves as signadot local connect --wait connect.
  • When connected to an operator version ≥ v0.16.0, traffic routing uses virtual IPs, which is safer and allows use in more contexts.

Deprecated

  • The outbound field of a local connection is only pertinent when connecting to Signadot Operator < v0.16.0.

2024-03-01

API / Control Plane

Added

  • Support for RouteGroup Time to Live (TTL).
  • Support for a new proxy endpoint allowing proxy connections in the upcoming version of the CLI.

Fixed

  • Fixed Bug when accessing forbidden resources (previously returned unauthorized instead of forbidden).

Chrome Extension v1.1.0

Fixed

  • Fixed bug causing the extension to occasionally lose state and inject empty routing-key headers.

Dashboard

Added

  • Support for RouteGroup Time to Live (TTL).

2024-02-15

API / Control Plane

Added

  • Support for Role-Based Access Control (RBAC) with an admin and member roles. See docs for reference.

Fixed

  • Fetching resources associated with a disconnected cluster now returns 502 HTTP status code instead of 500.

Dashboard

Added

  • Support for Role-Based Access Control (RBAC).
    • Inviting users via the "Users" page now allows specifying role as admin or member.
    • The "Users" list allows admin users to change the role of any other users to member.
    • Check docs for reference.

2024-02-08

Dashboard

Fixed

Resolved bug that prevented users from accessing kubectl configuration for playground clusters.

2024-02-01

Dashboard

Changed

Minimum viewport width supported is now 800px instead of 1024px.

Fixed

Resolved bug that was allowing users to self-delete from the Dashboard.

Security

Critical upgrades to dependencies.

CLI v0.6.1

Added

Shortname aliases for sandbox (sb), routegroup (rg), cluster (cl), and resourceplugins (rp)

Fixed

While connected using signadot local connect, the CLI now will not add entries to /etc/hosts which contain an IP address outside of the ranges configured to route to the cluster. Instead, it will log a warning.

2024-01-24

API / Control Plane

Fixed

  • Resolved issue that caused new users to be unable to register using email / password based authentication.
  • Resolved issue that caused an HTTP 500 error when new users attempt to login when "Auto-Provision users" setting in SSO is disabled. This now returns an HTTP 403 instead.
  • Resolved issue where preview URLs in an invalid format would return an empty response. They now return HTTP 400.

2024-01-18

API / Control Plane

Fixed

  • Fixed bug where listing resources associated with a sandbox when it is being deleted returned 500.

Dashboard

Added

  • New Logs UI panel in the Sandbox Details page.
  • Added support for SSO in the Enterprise Plan.

Operator v0.15.0

Added

Changed

  • DevMesh sidecars use the new Routes API, offering a significant speedup in in-cluster routing propagation.
  • DevMesh annotation (sidecar.signadot.com/inject) has changed:
    • The value http and grpc now behave the same as true and inject the envoy-based sidecar introduced in v0.14.
    • "false" (as a yaml string) will disable injection.
    • Parse errors of annotation values now also default to the above sidecar, logging an error.
  • Local connect improvements
    • Improved the method used to find the cluster IP network range in the Kubernetes cluster used to set up local networking. If it fails it will fall back to estimation as done previously.
    • The tunnel-api no longer returns DNS entries to the CLI that are outside the specified CIDR range. As a result, /etc/hosts managed in the CLI will omit entries outside the range.

Deprecated

  • The environment variable $SIGNADOT_SANDBOX_ID in scripts in Resource Plugin steps is now deprecated, and the newly added $SIGNADOT_SANDBOX_ROUTING_KEY should be used in its place.

Removed

  • Images (execpod-sidecar) and CustomResourceDefinitions (signadotresources and signadotresourceplugins) which were associated with V1 Resources have been removed. These have been superseded by V2 ResourcePlugins.

Fixed

  • Fixed bug in DevMesh routing mode where workloads in sandboxes with no associated services were marked not ready when forked.
  • Fixed bug causing traffic to be dropped temporarily when workstations running local sandboxes reconnect due to stale DNS entries.
  • Fixed bug that caused forked and local workloads to flap between not-ready and ready during updates.

2024-01-10

API / Control Plane

Removed

  • Legacy (v1) resource plugin support. This has been superseded by the v2 Resource Plugins API.

Dashboard

Fixed

  • Error handling in the case of 404 errors.
  • Incomplete fetching of information when deep-linking to view a particular cluster's information.
  • Issue where some links appear without styling.

Removed

  • Ability to view Legacy (v1) Resource Plugins has been removed.

2023-12-20

  • API: Fixed issue with updates to sandboxes reporting ready incorrectly before the update is reflected in the cluster.
  • API: Hidden field spec.LocalMachineID is no longer returned as part of the Sandbox Specification.
  • Previews: Connection pooling is now turned off for Preview URL requests to prevent unnecessary caching of in-cluster endpoint addresses.

2023-11-30

info

Operator v0.14.1 is now available
CLI v0.6.0 is now available
Chrome Extension v1.0.0 is now available (beta)

Upgrade Considerations for Operator v0.14.1

If you are consuming metrics for the operator within your infrastructure, v0.14.1 has a breaking change in the endpoints and the metric names. See docs for more details.

  • Operator:
    • Metrics
      • Breaking change: the metrics endpoints & names have changed in v0.14.1. Refer to monitoring docs for details.
      • Added new metrics that track individual entities (sandboxes, routegroups, resources), as well as component specific metrics. Refer to docs for details.
    • Routing / DevMesh
      • Fixed low timeouts on HTTP and gRPC connections proxied via the DevMesh sidecar.
      • Proxying of HTTP1.0 connections is now supported.
    • Routing / Istio
      • Fixed routing between local workloads and forked workloads in Istio mode. Note that this in turn disables the audit logging sidecar, thereby allowing the istio-sidecar to come up on the tunnel-proxy.
    • Local
      • Fixed bug that caused signadot local connections to not self-heal in certain cases when the connection is interrupted.
    • Misc
      • Improved naming of workloads created by sandboxes to accommodate more information about the baseline workload. Previously, both sandboxID and sandboxName were included in the workload name, of which sandboxID has been dropped in v0.14.1.
      • General improvements, and updates to Kubernetes dependencies (controller-runtime).
  • API:
    • Sandbox Updates: forks & local workloads can now be added / removed as part of sandbox updates with operator v0.14+
    • signadot/community#40: Support for hybrid sandboxes: You can now create sandboxes that have both forks and locals alongside each other with operator v0.14+.
  • CLI:
    • CLI v0.6.0 is now available.
    • In combination with v0.14.1 for the operator, sandboxes with local workloads now have machine affinity. When signadot local connect is run, any sandboxes that were previously associated with a machine and not deleted will re-associate with the machine. Refer to local workloads documentation for details.
    • --clean-local-sandboxes flag added to signadot local disconnect to delete the locally associated sandboxes during disconnect.
  • Other:
    • Chrome Extension for Signadot to set routing headers for Sandboxes and RouteGroups is now in beta.

2023-11-15

  • Integrations: Fixed bug where sandbox update fails validation but updates Pull Request commit status.
  • CLI: v0.5.4 is now available:
    • Fix to update /etc/hosts correctly when signadot local is running in docker.

2023-11-06

  • Web: Sandbox Details page is now updated with improved information architecture.
  • Web: Sandbox Details > Explore has been introduced for exploratory testing of HTTP APIs.

2023-11-02

  • API: Multiple ports can now be mapped for local workloads with Operator v0.14+.
  • Control Plane: Robustness improvements to the controller responsible for managing Sandboxes and RouteGroups within connected clusters.

2023-10-26

  • Integrations: GitHub Pull Request Integration now adds a status check to the Pull Request with information about the sandbox.
  • Control Plane: Robustness improvements & bug fixes to the controller responsible for managing Sandboxes and RouteGroups within connected clusters.

2023-10-12

info

Operator v0.14.0 is now available

Upgrade Considerations for Operator v0.14.0

Istio Routing

If you are not using Istio-based routing with Sandboxes, you can ignore this note. If you are using Istio for routing, the release v0.14 and above of the operator will require setting the istio.enabled parameter in helm values.

Note that this was automatically and implicit set in previous versions of the operator, and you will need to explicitly set this value to true during the upgrade to avoid disruption.

Removal of support for v1.createWorkspace API

The v1.createWorkspace API is no longer supported with operator v0.14 and above. Please use the stable v2 APIs instead.

  • Operator: New routing implementation addresses signadot/community#26.
    • Sandbox routing now supports pods serving on multiple ports.
    • Sandbox routing makes use of Kubernetes Services for routing and removes dependency on pod's ContainerPort specification.
      • Every sandboxed workload will now create Kubernetes Services that are counterparts of each baseline service and use them for routing.
    • Istio specific improvements:
      • Status reporting of VirtualServices being patched and partially managed by Signadot.
      • Error reporting if spec.gateways is specified in a VirtualService without a mesh entry. See docs for more details.
      • Various error reporting and debuggability improvements.
    • DevMesh (previously Signadot Sidecar Routing) specific improvements:
      • Updated sidecar webhook to avoid system namespaces (kube-system, kube-node-lease) to prevent warnings in GKE.
      • Reporting of an error if DevMesh routing is enabled but the baseline workload does not have the DevMesh sidecar injected.
    • Workload Management:
      • Sandbox deletions are faster because deletions of dependent resources are now performed in the background.
    • See architecture documentation for details on new Custom Resources and Controllers.
  • API: Improved validation when specifying Preview Endpoint targets.
  • UI: Logs Viewer now allows choosing which container to get logs from.
  • UI: Fix for list of Preview Endpoint targets showing too many entries at once when creating / editing a RouteGroup.
  • CLI: v0.5.3 is now available:
    • Fixed issue in signadot local connect that caused it to fail when using Azure AD legacy authentication in kubeconfig.

2023-09-27

  • API: Improved validation for custom patch in Sandbox fork customizations.
  • UI: Fixed bugs causing page crashes in the case of API errors.
  • CLI v0.5.2 is now available:

2023-09-16

  • UI: Dashboard redesign with several improvements to presentation and navigation.
  • API: Fixed bug where specifying an empty value for a fork environment variable returned an unexpected validation error.

2023-09-08

info

Operator v0.13.3 is now available

  • Operator: New implementation for Signadot Sidecar based on Envoy is now available. You can opt into this now if you use sidecar.signadot.com/inject: "true" in baseline manifests. This will be the default in v0.14+. If you are using Signadot Sidecar for routing, take a look at the Upgrade Considerations note below.
  • Operator: Fixed bug in Resource Controller that caused a resource to be marked as failed despite succeeding when pulling images took too long.
  • Operator: Fixed issue with iptables dependency that caused Tunnel Proxy to not initialize correctly. This bug affected running local workloads as part of sandboxes when deployed on recent Kubernetes Node OS versions.
Upgrade Considerations for Operator v0.13.3

This note only applies if you are using Signadot Sidecar-based routing, and using docker image mirroring. The image customization parameters have the following changes:

  • routeInit.image is now renamed to routeInit.legacy.image
  • routeSidecar.image is now renamed to routeSidecar.legacy.image

This has been done to make way for a new opt-in envoy-based sidecar implementation that will become the new default starting in v0.14+. When upgrading, if you are using either of the above helm parameters, please rename them to the new parameter names to avoid disruption.

2023-08-24

  • CLI v0.5.1 released, adding support for --wait and --wait-timeout to signadot local connect.
  • Sandbox apply operations with no change in spec (update scenario) are deemed no-op and are not processed.
  • Sandboxes now support new TTL with respect to updatedAt.
  • UI: Displays spec for Sandboxes, Resource Plugins and RouteGroups on the respective detail pages.

2023-08-02

info

Operator v0.13.2 is now available

  • Operator v0.13.2
    • Upgrades build toolchain to use go 1.20.
    • Improvements to helm templating for compatibility with versions of Helm <v3.10.
  • Playground clusters now support Sandboxes with local workloads.
  • UI: Minor presentation improvements.

2023-07-19

info

Operator v0.13.1 is now available

  • CLI v0.5.0 is now available, with support for local workloads in sandboxes (beta).
  • Operator
    • Local workloads: fix for slow closing of SSH-based inbound tunnels during graceful teardown.
    • Local workloads: fix for tunnel connection state not being preserved between compatible sandbox spec updates.
    • Local workloads: fixed bug in gRPC route-to-local when using Istio.
  • UI: Local workload in sandbox details UI now shows the connected username@hostname.

2023-07-13

info

Operator v0.13.0 is now available

  • Operator adds beta support for running sandboxed workloads locally on developer workstations.
    • Running local workloads in a sandbox requires CLI v0.5.0+ and supports MacOS & Linux.
    • New ExternalWorkload CRD and controller to enable core functionality.
    • Added tunnel-api, tunnel-proxy and auditor components (helm chart) to handle communication between workstations and the cluster.
  • UI: Improved grouping and aesthetic elements of sandbox details page for improved usability
  • UI: Fixed bug in status reporting for clusters where they would show an error state before any connection.

2023-06-21

  • Fixed RouteGroup Sandbox matching for Sandboxes using PR Integration. These sandboxes were previously not matched by RouteGroups.
  • General improvements and bug fixes.

2023-05-26

  • The UI now supports creating & updating sandboxes using a YAML specification
  • Improvements to onboarding flow using playground "Kubernetes clusters"
  • Security updates & bug fixes

2023-04-19

  • Support for provisioning a "playground" Kubernetes cluster on-demand for new users via the Dashboard.
    • The playground cluster will contain the Signadot Operator pre-installed.
    • The playgrund cluster will also contain a demo application http://github.com/signadot/hotrod.
    • Read-only kubectl access is also provided to this cluster via the UI.

2023-04-05

  • UI: Fixed bug where deprovision logs associated with a resource plugin were shown incorrectly.
  • UI: Improvements to presentation of details associated with resource plugins.

2023-03-30

  • New Resource Plugin framework is now available for creating and managing stateful resources with Sandboxes.
    • CLI has been updated to v0.4.0 and now supports resourceplugin subcommand.
    • Dashboard UI support for new resource plugins, resources and displaying logs associated with them.
    • Resource Plugins API is now available.
    • Resource Plugin Examples and Documentation

2023-03-22

info

Operator v0.12.0 is now available

  • Operator-level support for a new resource management framework has been added. This framework will make it easier to write, deploy and manage resources in sandboxes.

2023-02-03

  • GitHub Pull Request Integration is now Generally Available.
  • Dashboard changes to auto-suggest targets in a Route Group.

2023-01-25

  • RouteGroups which match a given Sandbox are now visible on the UI under the Sandbox Details page.
  • Removal: old cluster agents at version Signadot Operator v0.7.0 and lower will no longer be able to connect.
  • General improvements and bug fixes.

2023-01-18

  • Improved sandbox deletion cluster synchronisation performance.
  • Periodic cluster synchronisation of Sandboxes now respects force deletion.
  • General improvements and bug fixes.

2023-01-10

  • Fixed regression in routing behavior introduced by using multi-value headers in baggage and tracestate when setting context on preview requests.

2023-01-04

info

Operator v0.11.1 is now available

  • Fixed bug with status reporting for Route Groups when one or more sandboxes are not found.
  • Fixed bug in enforcement of lower-case names for Route Groups in the API.
  • Fixed bug in inference of containerPort on fork when not specified on baseline workload.
  • Fixed a bug in enforcement of fork conflicts in routing contexts upon creation of a Route Group.

2022-12-23

  • Route Groups are now generally available via the API and UI
  • CLI v0.3.7 is now available with support for Route Groups

2022-12-07

info

Operator v0.11.0 is now available

  • The operator now supports a new CRD of kind SignadotRouteGroup, used for routing requests through a set of sandboxes.
  • The operator has addressed the security issue USN-5710-1.

2022-11-17

  • Sandbox tags have been renamed to labels and made mutable.
  • Fixed bug with additional preview endpoints on the Dashboard.
  • General improvements and bug fixes.

2022-11-09

  • Sandboxes now support the downward api.
  • General improvements and bug fixes.

2022-11-02

info

Operator v0.10.1 is now available

  • Self-service upgrades to the team tier are supported via the Dashboard.
  • Sandbox TTL if set is visible on the Dashboard.
  • General improvements & bug fixes in the operator.

2022-10-19

  • Fixed misleading error message when creating a sandbox on a cluster before it is connected successfully.
  • Resource Plugin for Amazon S3 is now available.

2022-10-11

  • Fixed bug causing changes to preserved labels to not be propagated until a sandbox is updated.
  • Fixed bug in the dashboard that caused a 4xx error in analytics pages in some cases.
  • General improvements.

2022-10-03

  • Sandboxes now support TTL. You can now set a deadline (ttl duration) on a sandbox for it to be automatically cleaned up after the specified duration elapses.
  • Various bug fixes and improvements.

2022-09-15

info

Operator v0.10.0 is now available

  • Updates to forks and endpoints in sandboxes are now supported when this version of the operator or higher is installed.
  • CLI default timeout for waiting for readiness changed from 5 minutes to 3 minutes.
  • General improvements.

2022-08-08

  • API server now correctly lists host endpoints associated with a sandbox.
  • Dashboard analytics page no longer shows an error in case no clusters are connected.
  • General improvements.

2022-07-25

  • API server now refers to the routeType of a sandbox static endpoint as "host" instead of "static".
  • API server now enforces HTTP rate limiting when communicating with a cluster's Kubernetes API.

2022-07-19

info

Operator v0.9.9 is now available

  • A new API version (/api/v2) is now available, and new SDK versions will soon be released that use the new V2 API. Old code that uses the V1 API (/api/v1) will continue to work, but some code changes may be required when upgrading to build against the latest SDK versions.
  • Fixed a bug where the operator might make unnecessary cosmetic changes (e.g. "20ms" -> "0.02s") in Istio VirtualService routes that it didn't add.

2022-07-11

  • Analytics is now available for sandbox creation events via the Signadot Dashboard.

2022-06-22

  • Fixed bug with the handling of X-Forwarded-Host header when passed to Sandbox URLs. This header if set is now passed through correctly to the workload.
  • Fixed bug causing incorrect last updated time for sandboxes on the Dashboard.

2022-06-14

info

Operator v0.9.8 is now available

  • Creating Sandboxes is no longer supported via the UI. Please use the CLI or SDKs moving forward.
  • Fixed an issue with istio header routing to support partial matches on the baggage header value.
  • Improvements to Dashboard UI for configuring and setting up CLI.

2022-06-06

info

Operator v0.9.7 is now available

  • Resource provision/deprovision status is now shown in the Dashboard UI.
  • Resource plugins now support Image Spec Version 1.0.
  • Fixed bug in Operator that caused sandbox creation to fail when duplicate ContainerPorts are specified in the baseline deployment.

2022-06-01

info

Operator v0.9.6 is now available

  • Fixed bug where malformed image replacement rules caused a crash in the operator.

2022-05-31

info

Operator v0.9.5 is now available

  • Fixed an operator bug that could result in stale sandbox status information.
  • API and UI now support force-delete of sandboxes.
  • Various UI improvements related to Resources and Resource Plugins.

2022-05-16

info

Operator v0.9.4 is now available

  • Fixed CRD validation error when using Signadot Resources on a Kubernetes v1.19 or older cluster.

2022-05-04

info

Operator v0.9.3 is now available

  • Signadot Operator's controller-manager component now defaults to 512M of memory as opposed to 128M before.

  • Signadot Operator now allows customising memory and cpu resources via helm values.yaml, for example:

    operator:
    cpu: 100m
    memory: 512Mi
  • General improvements and bug fixes.

2022-04-26

info

Operator v0.9.2 is now available

  • Fix a crash in the UI which could occur while viewing Resource Details.
  • Signadot Operator now ignores image replacements on baseline deployments with a malformed container image specification

2022-04-25

info

Operator v0.9.1 is now available

  • Forks can now have dynamic environment variables that resolve to a reference to a different fork within the same sandbox.
  • Signadot Operator now allows overriding all images using helm value parameters.

2022-04-12

info

Operator v0.9.0 is now available

  • Sandboxes can now have associated Resources like databases whose lifecycles are tied to the Sandbox. Resources are provisioned and deprovisioned by Resource Plugins, which run inside the same cluster as the Sandbox.
  • Signadot Operator can now be granted permissions to specific namespaces instead of cluster-wide.

2022-04-05

  • Email / Password based authentication is now supported.
  • Inviting a new member to your organization now notifies them via email.

Bug Fixes:

  • Fixes to Dashboard UI.
  • Fixed Sandbox URLs not supporting the HTTP PATCH method.

2022-03-28

  • New "Authorized Domains" feature added that allows specific domains to be set up to automatically create users for people who log in with an email address that uses one of those chosen domains. This is a convenience feature for organizations with many users, so they do not have to manually create invitations for every person.

2022-03-21

info

Operator v0.8.1 is now available

  • Users can now authenticate using Google Sign In.
  • New user management capabilities, including invites.
  • Operator supports mode without mutating webhook.
  • Cluster Status now shows unhealthy for clusters awaiting their first connection.

2022-02-23

info

Operator v0.8.0 is now available

The major change in this release is that the "Workspace" concept has been renamed to "Sandbox" to better convey what it represents.

As part of the rename, we have introduced a new Helm chart (signadot/operator) and new container image repositories. The old Helm chart (signadot/workspaces) is still available, but is deprecated. Please see the Action Required note below if you need to migrate a cluster that was already running the old chart.

Note that the new Helm chart does not include the signadot Namespace object, so you will need to create the signadot namespace before installing the chart.

Action Required

Our Operator has been migrated to a new Helm chart. If you have version v0.7.x or below installed, you will need to fully uninstall the old signadot/workspaces chart before installing the new signadot/operator chart.\n\nNote that your forked workloads will be undeployed when you uninstall the old chart, but they will be redeployed automatically after installing the new chart and reconnecting the cluster to Signadot. Baseline workloads will not be affected by the migration.\n\nWhen you uninstall the old chart, the Kubernetes Secret object containing the cluster registration token will be lost. You can either generate a new cluster token (tied to the same cluster) in the Signadot Dashboard, or use a procedure like the following to save and restore the Secret:

# Save the cluster-agent Secret.
kubectl -n signadot get secret cluster-agent -o yaml > /tmp/cluster-agent.yaml

# Remove the old chart.
helm uninstall signadot-workspaces

# Wait for the old 'signadot' namespace to finish finalizing.
# Let this run until you see the message: namespaces "signadot" not found
while kubectl get ns signadot; do sleep 5; done

# Create the 'signadot' namespace again.
kubectl create ns signadot

# Repopulate the cluster-agent Secret.
kubectl -n signadot create -f /tmp/cluster-agent.yaml
rm /tmp/cluster-agent.yaml

# Install the new chart (note the different chart name).
helm repo update
helm install signadot-operator signadot/operator

2022-02-14

  • Dashboard: new overview page provides high level usage and statistics.
  • API/Operator: added support for forking Argo Rollouts into Signadot Workspaces.
  • Minor bug fixes and general improvements.

2022-02-07

info

Operator v0.7.16 is now available

  • Dashboard: Support for creating workspaces with deployments from multiple namespaces.
  • Minor bug fixes and general improvements

2022-01-31

info

Operator v0.7.15 is now available

  • Minor bug fixes and general improvements

2022-01-18

info

Operator v0.7.14 is now available

  • Dashboard: You can now see if a Cluster is connected and the Signadot components are running.
  • Minor bug fixes and general improvements

2022-01-10

info

Operator v0.7.13 is now available

  • Minor bug fixes and general improvements

2021-12-27

info

Operator v0.7.12 is now available

Major Updates

  • Operator: All operator components will be deployed into a single signadot namespace moving forward and the old signadot-operator namespace will be removed automatically. No additional action is required during an upgrade.

Bug Fixes

  • API: Fixed a spurious warning related to preview endpoints even when specified correctly.

2021-12-13

info

Operator v0.7.10 is now available

Major Updates

  • API: Added a new {workspaceID}/ready subresource that can be polled to wait for a workspace to become ready.

Minor Updates

  • API: Preview endpoints are returned as part of each Workspace result.

Bug Fixes

  • Dashboard: Fixed various UI bugs.

2021-12-06

info

Operator v0.7.9 is now available

Major Updates

  • Dashboard: Preview URLs are now generated for the baseline deployments shown on the UI so you can compare the forked Deployment's behavior to it.
  • Dashboard: Deployment details now shows additional tab for Pods that can be used to debug unhealthy workspaces.

Minor Updates

  • API: Custom preview endpoints can now have arbitrary top-level domain and point to ingress gateways and external endpoints, instead of being limited to internal Kubernetes services.
  • API: Workspaces are no longer created for API calls corresponding to GitHub Pull Requests that have already been closed.
  • Dashboard: Added examples for interacting with Workspaces via their preview endpoints.

Bug Fixes

  • Dashboard: Fixed crash when accessing workspace details for certain older workspaces with no name set.
  • API: Improved error codes and messages returned by workspace APIs.

2021-11-22

info

Operator v0.7.8 is now available

Minor Updates

  • Operator: Updated RBAC to add permission to read / write Kubernetes events for error reporting.
  • Dashboard: Workspace details page will report health of forked deployments.

Bug Fixes

  • Dashboard: Fixed an issue where custom environment variables disappear from the UI.
  • Dashboard: Fixed crash when visiting a deep link to the deployment details page.

2021-11-15

info

Operator v0.7.7 is now available

Major Updates

  • Operator: Workspace readiness is now reported by the operator via the SignadotWorkspace object.

Minor Updates

  • Dashboard: Each forked deployment will now report readiness status on the UI.
  • Dashboard: You will now get notified on the Clusters tab when there is a new operator version available.
  • Previews: Improved error messages when preview endpoint is configured incorrectly.

Bug Fixes

  • Operator: Fixed bug causing gRPC connection leaks in the sidecar proxy.

2021-11-08

info

Operator v0.7.6 is now available

Major Updates

  • Operator: Operator supports setting custom annotations on all components via helm values.

Minor Updates

  • Dashboard: Improved new user UX around connecting clusters and creating the first workspace.
  • Operator: Operator will gracefully reconnect when connection fails rather than crashing the agent pod.

Bug Fixes

  • API: Fixed bug where preview URLs associated with old workspaces were returned as null from the API.
  • Dashboard: Fixed bug where the name and tag of the custom image associated with a workspace created by a pull request was hidden.

2021-11-01

info

Operator v0.7.5 is now available

Major Updates

Minor Updates

  • Operator: RBAC permission for getting namespaces in connected clusters

Bug Fixes

  • Dashboard: Fixed crash in the environment variable editor

2021-10-21

info

Operator v0.7.4 is now available

Major Updates

  • Dashboard: New UI for modifying environment variables associated with any forked deployment at runtime.

Minor Updates

  • Dashboard: Performance and minor UX improvements.

Bug Fixes

2021-10-14

info

Operator v0.7.3 is now available

Major Updates

  • Dashboard: You can now see all forked deployments and images associated with them under workspace details.

Minor Updates

  • Dashboard: Improvements to Workspace Details and Create Workspace.

Bug Fixes

  • Operator: Fixed noisy logs.

2021-10-08

info

Operator v0.7.2 is now available

Bug Fixes

  • Operator: Fix Workspace creation error when one or more Kubernetes Services has an empty selector.

2021-10-07

info

Operator v0.7.1 is now available

Major Updates

  • Dashboard: You can now delete cluster entries that are no longer needed. This will also remove workspace entries tied to those clusters.

Minor Updates

  • Operator: The workspace copy of a Deployment will now always have only 1 replica.
  • Operator: The outgoing connection to the tunnel server is now made on the standard HTTPS port (443).

2021-09-30

info

Operator v0.7.0 is now available

Action Required

Before upgrading each cluster from v0.6.x or lower, you must first clear out the old SignadotWorkspace CRD:

kubectl delete crd signadotworkspaces.signadot.com

This will remove all the Workspaces temporarily, but will not affect the baseline services. You can then proceed to the normal upgrade process, after which the Workspaces in that cluster will be automatically restored. This is a one-time migration that will not be necessary for subsequent upgrades.

If you attempt the helm upgrade before running the above command, it will fail with the error \"Cluster\": field is immutable. If that happens, you can run the above command and then re-run the Helm upgrade. However, since the update may have been partially applied, you should also restart the operator's controller-manager afterward:

kubectl -n signadot-operator delete pods --all --wait=false

Major Updates

  • Operator: The SignadotWorkspace CRD is now a cluster-wide resource instead of being scoped to each namespace.
  • Operator: Workspace changes made through the API or Dashboard will now be re-synchronized into clusters periodically, allowing cluster state to self-heal after temporary synchronization errors.
  • Operator: Kubernetes resources like Services, Deployments, and Pods that the operator creates now have more descriptive object names.
  • Dashboard: The status of each Workspace is now shown in the table view, and any synchronization errors are shown on each Workspace's details page.
  • Dashboard: Table views for Workspaces, Clusters, and API Keys are now paginated and searchable.
  • Dashboard: You can now specify labels to copy from the baseline version of a Pod to the new Pod created for a workspace. This is in the global Settings page and applies to all workspaces in all clusters.

Minor Updates

  • Dashboard: The full value of an API key or Cluster Token is now only shown at creation time. If you don't have the value of an old token saved, you can delete it and create a new one.
  • Operator: You can now specify additional labels to be added to all the Signadot Operator components when installing the Helm chart. These labels will not apply to the copies created from baseline services.
  • API: Published "Create a new workspace" API for creating workspaces outside a PR workflow

Bug Fixes

  • Previews to gRPC services now work with more gRPC clients thanks to ALPN support.

2021-09-10

info

Operator v0.6.6 is now available

Major Updates

  • You can now use workspace preview URLs with services in Kubernetes that are serving HTTPS & gRPC traffic

2021-09-03

Major Updates

  • You can now preview Kubernetes manifest changes as well as code changes.
  • You can now connect multiple Kubernetes clusters to one Signadot account.

Minor Updates

  • Signadot now posts commit statuses on PRs to expose the workspace dashboard and preview URLs, instead of relying on your own CI automation to do it.

Bug Fixes

  • API: Fixed error when creating a workspace for a PR when there is already a workspace for another PR in the same repo.
  • Operator: Fix SchemaError ("array should have exactly one sub-item") when running on Kubernetes versions older than v1.18.

2021-08-25

info

Operator v0.6.4 is now available

Bug Fixes

  • Preview URLs: Fixed HTTP 405 error when using HTTP methods other than GET.
  • API: Fixed error when using different capitalization for GitHub org names, which should be case-insensitive.
  • Operator: Report an error in SignadotWorkspace status if a container image replacement was specified, but the intended image was not found in any baseline services.
  • Agent: Attempt to improve stability of tunnel connection and add logging for further debugging.

2021-08-20

info

Operator v0.6.3 is now available

Major Updates

Minor Updates

  • Use a custom header (signadot-api-key) for Signadot API Key authentication to avoid overlapping with headers used by the service being proxied through a preview URL.
  • Set the de facto standard X-Forwarded-Host and X-Forwarded-Proto headers on requests coming through a preview URL so the underlying service can discover its own preview URL for use in third-party callbacks.

Bug Fixes

  • Fix duplicate port name error when the baseline Deployment's Pod template has multiple container ports with the same name.

2021-08-10

info

Operator v0.6.2 is now available

Major Updates

  • Support previewing Kubernetes manifest changes (e.g. Deployment spec) from a PR branch, alone or alongside code changes.
  • Support setting ephemeral environment variable overrides (without touching Kubernetes manifests) from the workspace dashboard UI.
  • Create a workspace by hand through the web UI instead of only through a CI pipeline.

Bug Fixes

  • The Signadot GitHub App can now be installed on a personal (non-org) account.
  • The in-cluster Signadot Agent now uses keep-alive pings to prevent the tunnel connection from getting killed.

2021-07-22

info

Operator v0.6.1 is now available

Major Updates

  • Support Signadot API key auth for preview links in addition to GitHub OAuth sign-in, making it easier to access preview links programmatically or from the command line.
  • Signadot Operator no longer depends on cert-manager as a prerequisite.

2021-07-13

info

Operator v0.6.0 is now available

Major Updates

  • First Signadot release with support for Workspaces.