Summary: 81 instances, 71 unique

Text	Count
// TODO (shance) convert to composite types	1
// TODO: does not handle protocol?	1
Name: "gce_api_request_duration_seconds", // TODO: (shance) reconcile with cloudprovider	1
// TODO: (shance) refactor all scope to be above the serviceport level	1
// TODO: lock per service instead of global lock	1
// TODO: Change group name to discovery.k8s.io after approved	1
// TODO: Should we check name?  (In theory, an LB could expose it)	1
// TODO: (shance) refactor scope to be per-resource	1
// TODO: also validate ServicePorts in the exposed NEG annotation via webhook	1
// TODO: Remove this after deprecation	1
// TODO: Watch secrets	1
// TODO: Change group name to discovery.k8s.io after KEP is approved	1
// TODO: Implement ping for controller	1
// TODO: Need to skip processing if only Heartbeat/Ping timestamp changes, but the status does not.	1
// TODO: (shance) Replace this with data gathered from meta.AllServices	1
// TODO: Implement all the other store methods and make this a write	1
// TODO: Currently only used in testing. Modify so we don't leak compute	1
// TODO: fix this	2
// TODO: evolve as required, it's currently needed just to satisfy the	1
// TODO: Should this be math.MaxInt64?	1
// TODO: Set utilization and connection limits when we accept them	1
// TODO: why do we bail with success out if there is a permission error???	1
// TODO: for consistency with the rest of this sub-package this method	1
// TODO: We really want to know *when* a backend flipped states.	1
// TODO: does the path need to be different for each upgrade	1
// TODO: This check and EndpointsData.Ready field may be deleted once Endpoints support is removed.	1
// TODO: this should really be "fr" and "frs".	1
// TODO: (shance) refactor all scope to be above the serviceport level	2
// TODO: Include port, ip in the status, since it's in the health info.	1
// TODO: Better retries than relying on restartPolicy.	1
// TODO: Refactor for modularity	1
// TODO: move this logic into code gen	1
// TODO: Consider Ready conditions of the workloads before putting them into the EndpointSlice.	1
// TODO: avoid race condition here	2
// TODO: (shance) add unit tests	1
// TODO return error here	1
// TODO: The following options are considered:	1
// TODO (shance) Replace this with data gathered from meta.AllServices or discovery doc	1
// TODO: Make this more resilient to label changes by listing	1
// TODO: (shance) refactor to use filter	1
// TODO: (shance) implement zonal	1
// TODO: Refactor NEG to use cloud mocks so ctx.Cloud can be referenced within NewController.	1
// TODO: (shance) this file should ideally be combined with backends/features	1
// TODO: filter for the region the cluster is in.	1
// TODO: Replace Condition with standard Condition	1
// TODO: Use selector instaed of name to select endpointslices	1
// TODO: Find a better way to compare these	1
Name: "gce_api_request_errors", // TODO: (shance) reconcile with cloudprovider	1
// TODO: rename backend-service and health-check to not use port as key	1
// TODO: Leader Elect and Health Check?	1
// TODO: this path should be removed when external certificate managers migrate to a better solution.	1
// TODO: add link to Expose NEG documentation when complete	1
// TODO enhance this by looking at some metric from service controller to ensure it is up.	1
// TODO: revendor the GCE API go client so that this error will not be hit.	1
// TODO: (shance) convert this to use composite types	1
// TODO: Only lock per resource	1
// TODO: move event recorder initializer to main. Reuse it among controllers.	1
// TODO: If the port range and protocol don't match, recreate the rule	1
// TODO: Handle the last case better.	1
// TODO: Use standardized conditions (KEP-1623)	1
// TODO: (shance) refactor this so we don't need an empty arg	2
// TODO: (shance) refactor scope to be per-resource	2
// TODO: make AggregateList return map[meta.Key]Object	1
// TODO: this is copied from "cloud.google.com/go/compute/metadata"	1
// TODO: what if the port changes?	1
// TODO: this handler exists only to avoid breaking existing deployments	1
// TODO: Pipe the clusterName through, for now it saves code churn	1
// TODO: Refactor for modulatiry.	1
// TODO: This should be a list, not a map so the generated code ordering is stable	1
// TODO remove the 2 definitions once they are added in legacy-cloud-providers/gce	1
// TODO extend this to unknown resources with a client pool	6