Developer Portal

  • What’s changed

    Changed

    • We added better error handing to the app version picker (#1266)
    • Show the spinner and disable the submit button when the request to update an app’s version is in flight. (#1283)
    • Make the clusters list responsive (#1254)
    • Give faster feedback when changing the cluster name by making it an optimistic update (#1248)

    Fixed

    • Fix an exception that could occur in the error fallback helper (#1253)
    • Fixed various exceptions that could occur while handling errors (#1249)
    • Fixed a bug when showing the app detail pane for apps on internal or otherwise hidden catalogs (#1252)
    • Fix admin login if switching endpoints (#1246)
    • Fix for undefined API endpoint in cluster details view (#1247)
    • Fixed a crash when viewing an organization that doesn’t exist by redirect to organizations home if the organization is not found (#1245)

    Dev changes

    • Refactor app catalog tests (#1251)
    • Create custom error object with detailed stack traces (#1250)
    • Enable whitelisting support for all providers by removing AWS condition from the whitelisting Ingress annotation (#1267)
  • Changed

    • The user-installed app list now auto refreshes (#1234)

    Fixed

    • We fixed an error that occurred when deleting an app. (#1235)
  • Added

    There have been some additions regarding the App Catalog and Apps in general:

    • We added a way to edit the version of an installed app. (#1227, #1212)
    • We’ve also added a version picker during app installation, so you can now pick the version of the app that you want to install (#1202)
    • There is now a table that shows chart versions and what app versions they will install when viewing an app from the App Catalog (#1200)
    • Add lastly, we’ve added kiam and external-dns to the list of preinstalled apps on clusters that indeed have these apps (#1197)

    We’ve also made upgrade availability information more visible in more places:

    • We’ve added an upgrade notice in the clusters list, so you can see at a glance which clusters are eligible for an upgrade (#1204)
    • Similar to the above, we’ve also added an upgrade available icon in the list of clusters on the organization detail page (#1208)

    Changed

    • Some modals could stay open even when changing routes. We now make sure modals close when the route changes (#1211)
    • We now ensure that the organization provider credentials form closes when you leave the page (#1209)

    Fixed

    • We fixed a memory leak in the app installation form (#1210)
    • We now ensure that the upgrade cluster modal does not re-render unnecessarily after dispatching an action (#1233)
    • Check cluster existence in dashboard resources component to avoid some potential error cases (#1230)
    • We fixed a bug where clusters would be created in the currently selected organization, instead of the organization you clicked “create cluster” in (#1207)
    • And lastly, we fixed an issue triggering the update cluster modal (#1206)

    Development changes / refactorings

    • Change appReducer to mainReducer (#1228)
    • Create custom response object (#1226)
    • Return empty array if no cluster in selectClusterNodePools (#1231)
    • Refactor keypair test and fix overloaded action type. (#1223)
    • Write tests and comments for the new model layer (#1203)
  • Fixed

    • Fixed an infinite loading spinner issue when opening Happa. (We had to add support for ‘giantswarm’ authorization schemes in a new client we introduced while refactoring) (#1195)

    Changed

    • Remove an exclamation mark from the app detail pane in the cluster detail page. (#1194)
  • Changes

    • Removed the “Browse Catalog” button and made the whole catalog repo clickable (#1177)
    • We now autofocus on the cluster name input in the cluster creation screen (#1172)

    Fixes

    • Fixed redirection behaviour when visiting a cluster that no longer exists (#1165)
    • We ‘accidentally a word’ on the cluster detail page in the custom credentials area (#1167)
    • Show app name instead of object Object in flash and error to reducer (#1171)
    • We now show an error when cluster creation fails (#1169)
    • Fixed an exception when clicking on the availability zone icons in the cluster details screen (#1151)

    Dev Changes / Refactorings

    • Create model layer for the control plane API (#1145)
    • In preparation of an upcoming feature we’ve added a version picker component, which is not yet implemented in its intended form. (#1164)
    • Added various integration tests
      • Clusters listing (#1176)
      • User management page and modals (User management is a Giant Swarm staff only feature) (#1168)
      • Reseting password (#1154)
    • Ellipsis in members list and as a composable (#1170)
  • Fixed

    • Fixed an infinite spinner on the app details page (#1150)

    Development changes

    • Change organization tests to not use persisted nock mocks. (#1146)
    • Add _REQUEST suffix to actions (#1149)
  • Added

    • In preparation of an upcoming Giant Swarm release, we have added support for picking Availability Zones when creating Azure clusters. (#899)

    Changed

    • We now wait till we are sure the cluster delete request succeeded before redirecting to the organization page (#1099)

    • We’ve added a “Nodes not ready yet” message to the node count when a cluster is new, instead of showing 0 Nodes which might be a bit confusing. (#1137, #1103, #1141, #1108, #1129, #1128)

    • Change the error message when Happa fails to fetch or process releases(#1127, #1122)

    Fixed

    • Temporary fix for last updated label (#1116)
    • Fixed a flash of content in the v4 cluster details views (#1143)
    • We fixed an authentication issue with the User invite feature (only available to Giant Swarm staff at the moment). It was using the old JWT token instead of a new one after the token got stale. (#1089)

    Dev changes / Refactorings

    • Added tests for app installation and editing (#1120)
    • Update production build (#1130)
    • Adapt Happa to use new error/loading reducers in clusterReducer and cluster views (#1109)
    • Write integration test for the OAuth admin login user flow (#1106)
    • Write integration test for the Sign Up user flow (#1102)
    • Change ErrorFallback so that it accepts an array (#1104)
    • Add error reducer (#1100)
    • Add entity error reducer and ErrorFallback component (#1101)
    • Entity loading reducer (#1125)
    • Remove deleted cluster from calls in batchedRefreshClusters (#1138)
    • Fix capabilities not being injected during the deployment process (#1148)
  • Bug Fixes

    • We now reset the create organization modal after submitting or closing it. (#1087)
    • We also fixed an issue with our CI now that we are prepending tags with v (#1088)
  • New Features

    • When viewing a cluster that does not come with a preinstalled ingress we now show a message explaining why and include a link to get more information. (#1079)

    Changes

    • We’ve made a change to avoid showing the upgrade available message for incompatible version jumps on AWS. (#1083)
    • The names of node pools are now truncated only if there is not enough space to display them (#1084)
    • The kubeconfigs generated when creating a keypair are now aligned with what gsctl produces. They now have cluster specific values for the name, username, and context fields (#1073)

    Bug Fixes

    • Improve how we deal with clusters that have been deleted. (#1086)
    • Fixed certain issues on the cluster detail page (#1082, #1062, #1059)
    • Fixed invite modal logic (#1072) (This is relevant only to Giant Swarm staff, customers are not able to invite users)
    • Reduce the number of api calls made when switching organizations (#1060)

    Dev changes / Refactorings

    • Produce node pools to selector (#1061)
    • Refactor OAuthCallback route into function component (#1076)

    Note: We’ve started adding v to our tags, because that is the recommended practice and how other repos throughout Giant Swarm do it too!

  • Changes

    • Instead of waiting for all requests to finish, we render the cluster overview immediately with some placeholders for things like CPU, RAM, Nodecount. (#1048, #1036)

    Bugfixes

    • Fixed a crash when trying to render clusters with an unknown instance type (#1058)
    • Corrected the endpoint used when upgrading our latest (v5) clusters (#1057)
    • Fixed various display issues in the cluster detail view (#1056, #1053, #1054, #1037)
    • Fixed an error after creating clusters with an older version (#1051)
    • Fixed a bug when deleting clusters that could cause it to reappear in a deleted state. (#1035)

    Dev Changes

    • Fixes and improvements in our testing

      • Added tests for components with copyable content (#1038)
      • Fixed memory leaks and errors in tests console (#1055)
      • Add integration tests and testUtils paths to TS config (#1050)
    • An internal refactoring to refer to our latest clusters as “v5 clusters” instead of nodepools clusters (#1047)

    • Dependency updates (#1046)

    • Moved all URLs to an enum-like data structure and provide helpers for easily manipulating them (#1031)