Detect when interface cast fails and return an error instead of panic.

[timflyio]

Change Summary

What and Why: Fix a panic

% flyctl ext k8s save-kubeconfig fks-tim-newsham-fks-sake
Oops, something went wrong! Could you try that again?

interface conversion: interface {} is nil, not string
goroutine 1 [running]:
runtime/debug.Stack()
    runtime/debug/stack.go:26 +0x64
github.com/superfly/flyctl/internal/sentry.printError({0x104f91b20, 0x140006646f0})
    github.com/superfly/flyctl/internal/sentry/sentry.go:190 +0x168
github.com/superfly/flyctl/internal/sentry.Recover({0x104f91b20, 0x140006646f0})
    github.com/superfly/flyctl/internal/sentry/sentry.go:180 +0xb8
main.run.func1()
    github.com/superfly/flyctl/main.go:40 +0x34
panic({0x104f91b20?, 0x140006646f0?})
    runtime/panic.go:785 +0x124
github.com/superfly/flyctl/internal/command/extensions/kubernetes.runSaveKubeconfig({0x105322348, 0x1400054d410})
    github.com/superfly/flyctl/internal/command/extensions/kubernetes/kubeconfig.go:45 +0x2d8
github.com/superfly/flyctl/internal/command/extensions/kubernetes.saveKubeconfig.New.newRunE.func2(0x140006c8008, {0x1400005aa00?, 0x4?, 0x1047e1859?})
    github.com/superfly/flyctl/internal/command/command.go:140 +0x188

probably related to a failure provisioning the cluster:

% flyctl ext k8s create -o personal -n fks-sake
Fly Kubernetes is in beta, it is not recommended for critical production use cases. For help or feedback, email us at [email protected]
Some regions require a Launch plan or higher (bom).
See https://fly.io/plans to set up a plan.

? Choose the primary region (can't be changed later) [staff] Ashburn, Virginia (US) (qmx)
Error: failed to create etcd cluster - Failed to create volume: Status 500, Trace ID: b273d47d095c8504f7f14e76c573850d, Body: {"error"=>"internal: failed to get app: sql: no rows in result set"}

How: by detecting when an interface cast fails and returning an error instead of crashing.

Related to:

---

Documentation

• Fresh Produce
• In superfly/docs, or asked for help from docs team
• n/a

[senyosimpson]

maybe we should add "provisioning your cluster may have failed" to it? only because the only way this would fail is if something got borked on our end and there's no real way to fix it by running it again

[timflyio]

I updated the error message and included a potential tip for fixing the situation (delete and reprovision).

[senyosimpson]

one comment but not important, lgtm