Improve k8s volume size calculations

[johnbradley]

PR #140 uses the job's volume size for the job-data, output-data, tmpout volume sizes.
The /tmp volume is hard coded to be 1 G.
Need to come up with more accurate sizes for these.

[johnbradley]

I had a job that failed due to /tmp filling up.
The workflow failed running gatk4 SortSam resulting in a No space left on device error .

Related code:

lando/lando/k8s/jobmanager.py

Line 10 in c6eab0b

TMP_VOLUME_SIZE_IN_G = 1

lando/lando/k8s/jobmanager.py

Lines 70 to 76 in c6eab0b

	def create_tmp_persistent_volume(self):
	self.cluster_api.create_persistent_volume_claim(
	self.names.tmp,
	storage_size_in_g=TMP_VOLUME_SIZE_IN_G,
	storage_class_name=self.storage_class_name,
	labels=self.default_metadata_labels,
	)