feat(greenhouse): adds dashboards for alerts and plugins#1737
feat(greenhouse): adds dashboards for alerts and plugins#1737
Conversation
This adds two dashboards: Alerts and Plugin. The main design drivers for both of these dashboards has been to (i) use the greenhouse_* metrics, (ii) be a 1-1 mapping of the Greenhouse Alerts, and (iii) give context to the alerts more than simply a number. The way it has been designed, to not make it completely unusable due to the number of params (e.g. clusterNames, nodes, jobs) is to add UX elements that allows for filtering some params. --------- Signed-off-by: Simon Olander <[email protected]>
|
#Plugins 
|
Alerts
|
(chore): retain catalog secret err in inventory status (chore): revert error propagation in source/objectReadiness (chore): reflect aggregated error in statusCondition (chore): format aggregated errs in catalog (chore): increase e2e timeout to 3m (chore): move error struct to source Signed-off-by: abhijith-darshan <[email protected]>
|
Hey Simon, thanks for driving this. For a few panel groups in the alerts dashboard I see some redundancy to existing dashboards. The Operator Alerts on the Alerts Plugin were modelled after the controller runtime metrics dashboard. Same for the Proxy alerts which are covered by the Proxy Overview. When writing the original issue I was thinking more in the direction of having an overview by Greenhouse resources/ group of resources. As an example for the Team Alerts. This could be it's own Organization dashboard. Showing the overall status of the organisation (e.g. The plugin dashboard is going into this direction. For the panels |
(chore): increase e2e timeout to 5m in actions (chore): e2e workflow - revert to main branch Signed-off-by: abhijith-darshan <[email protected]>
(chore): fix lint Signed-off-by: abhijith-darshan <[email protected]>
* Authorization webhook with insecure config (WIP) On-behalf-of: @SAP [email protected] * fix authz chart, setup insecure authorization webhook server On-behalf-of: @SAP [email protected] * sets the local image for the authz deployment On-behalf-of: @SAP [email protected] * apply authz service manifest in local setup, set up the service as NodePort On-behalf-of: @SAP [email protected] * mount custom resolv config for kube-apiserver in kind cluster to use CoreDNS, use ClusterIP service for authz On-behalf-of: @SAP [email protected] * enable secure mTLS connection to authorization webhook from apiserver On-behalf-of: @SAP [email protected] * handle authorization of requests for greenhouse resources based on owned-by label; add serviceaccount with rbac to authz On-behalf-of: @SAP [email protected] * add greenhouse prefix to authz resources, tidy up logs On-behalf-of: @SAP [email protected] * filter verbs in requests redirected to authorizer webhook On-behalf-of: @SAP [email protected] * update verbs in authz webhook matchConditions On-behalf-of: @SAP [email protected] * add authz certs generation to setup-manager target, remove custom resolv.conf and use NodePort service with loopback IP instead On-behalf-of: @SAP [email protected] * disable authz by default in greenhouse chart, revert chart versions, fix lint suggestions On-behalf-of: @SAP [email protected] * extract code to init func in authz On-behalf-of: @SAP [email protected] * fix helm lint On-behalf-of: @SAP [email protected] * use the client.Client with RESTMapper instead of dynamic client in authz On-behalf-of: @SAP [email protected] * put authz-certs target as dependency to make setup-manager On-behalf-of: @SAP [email protected] * template authz deployment and service values; replace authz modification in local setup with chart overrides; remove unnecessary match condition On-behalf-of: @SAP [email protected] * feat: add insecure option for running authz server, update chart On-behalf-of: @SAP [email protected] * set v1.33.4 for e2e test action, remove v1.32.8 from nightly e2e On-behalf-of: @SAP [email protected] * set v1.33.4 for e2e tests; fix unhandled error On-behalf-of: @SAP [email protected] * revert v1.32.8 version to e2e matrix for remote cluster On-behalf-of: @SAP [email protected] * change admin cluster config for e2e tests action to not include authorization webhook On-behalf-of: @SAP [email protected] * fixes after code review, change env to AUTHZ_TLS_DISABLED with default false, set insecure authz webhook for local setup On-behalf-of: @SAP [email protected]
Agreed, the other dashboards look a lot nicer. I deleted these duplicates.
I added a new Dashboard called "Organization". To me it is a nice start, but it looks a bit empty tbh
That sounds like a good idea to me as well. I started with the alert-angle to have something to start with, but I agree that it is redundant as you can render the exact alert in Prometheus/thanos. Only drawback with a TimeSeriesChart however is that it looks a bit cluttered. I agree that it gives an operator way more information than a simple "boolean"-table. But It is not super easy to make render it in a nice looking way. |
--------- Signed-off-by: Simon Olander <[email protected]>
those gno plugins are not part of obs-eu-nl-1 |
(chore): e2e test to verify label propagation (chore): catalog propagate only owned-by label (chore): remove extraction test cases (chore): check suspended status condition (chore): remove nested eventually Signed-off-by: abhijith-darshan <[email protected]>
Flux is deployed into the local environment by default Expressions can be used by default as well
they are according to the metrics 🤔 |
|
Handing over this PR to someone else for the time-being. |
* fix: set HelmReconcileFailedCondition if HelmRelease creation fails for Plugin On-behalf-of: @SAP [email protected] * fix flaky E2E test from catalog scenario On-behalf-of: @SAP [email protected]
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
This helper will update the version comment to use semver in case of just a major version being used
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
…very tool to helm for helm controller tests (#1772) On-behalf-of: @SAP [email protected]
currently the Plugin controller will set .spec.test.enable=false this is overwritten by flux and results in updates to the helmrelease
(chore): add PD and CPD API type methods (chore): add ready condition to printer columns remove description from printer column (chore): unified methods for PD and CPD controllers utils unifies setupManager with watches for both CPD and PD controllers Generic PluginDefinition interface for extracting spec, chart name, setting conditions, re-use in other controllers unifies helm repository and helm chart resource create / update for PD and CPD (chore): use unified methods in PD and CPD controllers (chore): update test cases (chore): refactor ready condition (chore): simplify method signature (chore): fix lint errors (chore): force reconcile until kustomization ready (chore): use update instead of patch for annotations (chore): remove nesting eventually (chore): rename enqueue func (chore): remove nested eventually (chore): apply suggestions from code review (chore): remove redundant predicateFunc (chore): remove SetUnknownCondition and rename ChartName() to FluxHelmChartResourceName() (chore): initialize conditions (chore): pass gomega to underlying helper (chore): use owns and controller reference on helm charts (chore): rename enqueue event handler method Signed-off-by: abhijith-darshan <[email protected]>
(chore): removes unused manifests (chore): adds license headers (chore): adds local registry ca mount to source-controller (chore): adds local registry repository const (chore): rename org name (chore): rebase Makefile Signed-off-by: abhijith-darshan <[email protected]>
(chore): use CreateOrPatch instead of update (chore): use slices sort Signed-off-by: abhijith-darshan <[email protected]>
* fix: predicates for Plugins in PluginPresetController On-behalf-of: @SAP [email protected] * move IgnoreDeletingResources to its own predicate func, fix linter suggestion On-behalf-of: @SAP [email protected] * remove unnecessary DeleteFunc line from predicate, add UID check to the test On-behalf-of: @SAP [email protected]
* fix(deps): update kubernetes packages * adapt webhooks to Generic Validator and Defaulter from controller-runtime v0.23.0 On-behalf-of: @SAP [email protected] * adapt events to the events API migration introduced by controller-runtime v0.23.0 On-behalf-of: @SAP [email protected] * upgrade mocks with mockery On-behalf-of: @SAP [email protected] * set events.k8s.io group to events On-behalf-of: @SAP [email protected] --------- Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> Co-authored-by: Zaggy21 <[email protected]>
* (chore): separate authz commands from root Makefile Signed-off-by: abhijith-darshan <[email protected]> * Apply suggestions from code review Co-authored-by: Krzysztof Zagorski <[email protected]> --------- Signed-off-by: abhijith-darshan <[email protected]> Co-authored-by: Krzysztof Zagorski <[email protected]>
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
* enclose oauth clients fetching in Eventually block, fix assertions On-behalf-of: @SAP [email protected] * fix flaky teamrolebinding test On-behalf-of: @SAP [email protected]
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
this bumps dexidp to the commit fixing the issue with net.URL > 1.25.2, 1.24.8 currently there is no new release this is to resolve security issues in the GHAS
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
github-actions
bot
added
size/L
documentation
IvoGoman
left a comment
IvoGoman
left a comment
There was a problem hiding this comment.
Thanks for the big rework @olandr 🎉
A couple of things which would be cool:
Organization
- the Variable Organization does not behave as intended. Can you use
greenhouse_organization_readyas the Series selector. (This only being greenhouse will be addressed soon) - Number of Team Members would be nice as an Timeseries Chart with
{{team}}in the legend
For the other two let's get them in use them and understand where we can tweak them.
7 participants
This adds two dashboards: Alerts and Plugin. The main design drivers for both of these dashboards has been to (i) use the greenhouse_* metrics, (ii) be a 1-1 mapping of the Greenhouse Alerts, and (iii) give context to the alerts more than simply a number.
The way it has been designed, to not make it completely unusable due to the number of params (e.g. clusterNames, nodes, jobs) is to add UX elements that allows for filtering some params.
Signed-off-by: Simon Olander [email protected]