fix(application): handle storage details conversion error

[anvial]

Description

This PR ensures that the application correctly handles storage conversion errors and retries appropriately until the storage is ready.

• Added error handling for "cannot convert storage details" in applicationsClient.ReadApplication.
• Improved retry logic for storage readiness in ReadApplicationWithRetryOnNotFound.

Add a fatal error check to the retry logic around reading an application. Fail if ApplicationsInfo fails in a way other than expected. See a "connection refused" error, which is intermittent, so allow for retry in that case.

Fixes: #520

Type of change

• Change in tests (one or several tests have been changed)
• Bug fix (non-breaking change which fixes an issue)

QA steps

Also run reproduction steps from bug.

terraform {
  required_providers {
    juju = {
      source  = "juju/juju"
      version = "0.14.0"
    }
  }
}
provider "juju" {}

resource "juju_model" "sunbeam" {
  name = var.model

  cloud {
    name   = var.cloud
    region = "localhost"
  }

  config = var.config
}

resource "juju_application" "mysql" {
  name  = "mysql"
  trust = true
  model = juju_model.sunbeam.name

  charm {
    name    = "mysql-k8s"
    channel = "8.0/stable"
    base    = "ubuntu@22.04"
  }

  storage_directives = { database = "1G" }
}

resource "juju_application" "mysql1" {
  name  = "mysql1"
  trust = true
  model = juju_model.sunbeam.name

  charm {
    name    = "mysql-k8s"
    channel = "8.0/stable"
    base    = "ubuntu@22.04"
  }

  storage_directives = { database = "1G" }
}

resource "juju_application" "mysql2" {
  name  = "mysql2"
  trust = true
  model = juju_model.sunbeam.name

  charm {
    name    = "mysql-k8s"
    channel = "8.0/stable"
    base    = "ubuntu@22.04"
  }

  storage_directives = { database = "1G" }
}

resource "juju_application" "mysql3" {
  name  = "mysql3"
  trust = true
  model = juju_model.sunbeam.name

  charm {
    name    = "mysql-k8s"
    channel = "8.0/stable"
    base    = "ubuntu@22.04"
  }

  storage_directives = { database = "1G" }
}

resource "juju_application" "mysql4" {
  name  = "mysql4"
  trust = true
  model = juju_model.sunbeam.name

  charm {
    name    = "mysql-k8s"
    channel = "8.0/stable"
    base    = "ubuntu@22.04"
  }

  storage_directives = { database = "1G" }
}

resource "juju_application" "mysql5" {
  name  = "mysql5"
  trust = true
  model = juju_model.sunbeam.name

  charm {
    name    = "mysql-k8s"
    channel = "8.0/stable"
    base    = "ubuntu@22.04"
  }

  storage_directives = { database = "1G" }
}

resource "juju_application" "mysql6" {
  name  = "mysql6"
  trust = true
  model = juju_model.sunbeam.name

  charm {
    name    = "mysql-k8s"
    channel = "8.0/stable"
    base    = "ubuntu@22.04"
  }

  storage_directives = { database = "1G" }
}

resource "juju_application" "mysql7" {
  name  = "mysql7"
  trust = true
  model = juju_model.sunbeam.name

  charm {
    name    = "mysql-k8s"
    channel = "8.0/stable"
    base    = "ubuntu@22.04"
  }

  storage_directives = { database = "1G" }
}

resource "juju_application" "mysql8" {
  name  = "mysql8"
  trust = true
  model = juju_model.sunbeam.name

  charm {
    name    = "mysql-k8s"
    channel = "8.0/stable"
    base    = "ubuntu@22.04"
  }

  storage_directives = { database = "1G" }
}

variable "model" {
  description = "Name of Juju model to use for deployment"
  default     = "sunbeam"
}

variable "cloud" {
  description = "Name of K8S cloud to use for deployment"
  default     = "microk8s"
}

variable "config" {
  description = "Set configuration on model"
  default     = { "workload-storage" : "microk8s-hostpath" }
}

rm -rf .terraform* && rm -rf terraform*

terraform init -upgrade  && terraform plan && terraform apply -auto-approve -parallelism=3

[cderici]

LGTM, QA went well, tried with various -parallelism values (including -parallelism=3) and was able to apply the plan each time 👍

[anvial]

Thx, I also double checked... I'll land these changes.

Heather, thx for catching IsFatalError case, I didn't see it on my local machine.