Use filtering to improve performance of agent pool data source

[JarrettSpiker]

Description

Once agent pool filtering is enabled (currently off in production behind the remove-agent-pool-limit feature flag) we can filter agent pool list requests by agent pool name. This improves performance of the agent pool data source when the organization has >20 agent pools by reducing the number of API calls we need to make in order to loop through all of the pages of agent pools

Testing plan

1. Set up a TFC organization with >20 agent pools, and a terraform config with a data source matching one of those agent pools
2. Verify that with the remove-agent-pool-limit feature flag off, the data source resolves the agent pool
3. Verify that with the remove-agent-pool-limit feature flag on, the data source resolves the agent pool even faster
   • The difference for me, against an org with ~1100 agent pools, is 6s w/ the feature flag vs. 1m12s w/o the feature flag

External links

• API documentation: In progress...
• go-tfe PR

Output from acceptance tests

 TESTARGS="-run  TestAccTFEAgentPoolDataSource_" make testacc
==> Checking that code complies with gofmt requirements...
TF_ACC=1 go test $(go list ./... |grep -v 'vendor') -v -run  TestAccTFEAgentPoolDataSource_ -timeout 15m
?   	github.com/hashicorp/terraform-provider-tfe	[no test files]
=== RUN   TestAccTFEAgentPoolDataSource_basic
--- PASS: TestAccTFEAgentPoolDataSource_basic (8.57s)
PASS
ok  	github.com/hashicorp/terraform-provider-tfe/tfe	8.798s
?   	github.com/hashicorp/terraform-provider-tfe/version	[no test files]

...

---

• To see the specific tasks where the Asana app for GitHub is being used, see below:
  • https://app.asana.com/0/0/1202293596899902

[sebasslash]

Looks 🔥 (won't approve until GA though)! Do you think we can improve the import of this resource with this new functionality?

[JarrettSpiker]

Looks 🔥 (won't approve until GA though)! Do you think we can improve the import of this resource with this new functionality?

Im not sure exactly what you mean? Is there inefficiencies in how resources are imported by default?

[sebasslash]

Not an inefficiency per-se, but if we look at the current import function:

terraform-provider-tfe/tfe/resource_tfe_agent_pool.go

Lines 17 to 19 in 6f515c3

	Importer: &schema.ResourceImporter{
	State: schema.ImportStatePassthrough,
	},

It just supports importing by ID. We can potentially extend this to support importing using an org-name/agent-pool-name pair. This might be way out of the scope for this PR come to think of it. Just something to consider!

[sebasslash]

Approved 👍