fix(GraphQL): fixes flaky test for subscriptions.

[JatinDev543]

Fixes DGRAPH-2305

---

This change is 

Docs Preview:

[arijitAD]

Reviewable status: 0 of 1 files reviewed, 1 unresolved discussion (waiting on @jatindevdg, @MichaelJCompton, and @pawanrawal)

---

graphql/e2e/subscription/subscription_test.go, line 584 at r1 (raw file):

	}
	addResult = add.ExecuteAsPost(t, graphQLEndpoint)
	require.Nil(t, addResult.Errors)

Why weren't we sending the request previously? Was it intentional or did we miss out?

[JatinDev543]

Reviewable status: 0 of 1 files reviewed, 1 unresolved discussion (waiting on @jatindevdg, @MichaelJCompton, and @pawanrawal)

graphql/e2e/subscription/subscription_test.go, line 584 at r1 (raw file):

	}
	addResult = add.ExecuteAsPost(t, graphQLEndpoint)
	require.Nil(t, addResult.Errors)

Why weren't we sending the request previously? Was it intentional or did we miss out?

we missed it somehow.

[pawanrawal]

Reviewable status: 0 of 1 files reviewed, 4 unresolved discussions (waiting on @arijitAD, @jatindevdg, @MichaelJCompton, and @pawanrawal)

---

graphql/e2e/subscription/subscription_test.go, line 754 at r2 (raw file):

	err = json.Unmarshal(res, &resp)
	require.NoError(t, err)
	require.Nil(t, resp.Errors)

Shouldn't we check resp.Errors before we do the json.Unmarshal?

---

graphql/e2e/subscription/subscription_test.go, line 768 at r2 (raw file):

	// add delay for 2nd subscription  to timeout
	// Wait for JWT to expire.
	time.Sleep(20 * time.Second)

Why does this need to be increased? Is 10 not enough?

---

graphql/e2e/subscription/subscription_test.go, line 785 at r2 (raw file):

	}

	addResult = add.ExecuteAsPost(t, graphQLEndpoint)

Didn't we fix this before? I thought we did.

[JatinDev543]

Reviewable status: 0 of 1 files reviewed, 4 unresolved discussions (waiting on @arijitAD, @MichaelJCompton, and @pawanrawal)

---

graphql/e2e/subscription/subscription_test.go, line 754 at r2 (raw file):

Previously, pawanrawal (Pawan Rawal) wrote…

Shouldn't we check resp.Errors before we do the json.Unmarshal?

no, i checked that, resp is filled while unmarshling, so we should check resp.errors after unmarshling.

---

graphql/e2e/subscription/subscription_test.go, line 768 at r2 (raw file):

Previously, pawanrawal (Pawan Rawal) wrote…

Why does this need to be increased? Is 10 not enough?

fixed it.

---

graphql/e2e/subscription/subscription_test.go, line 785 at r2 (raw file):

Previously, pawanrawal (Pawan Rawal) wrote…

Didn't we fix this before? I thought we did.

i missed it at some places.fixed it.

[codelingo]

LGTM. 3 rules errored during the review.

[codelingo]

1 issues found. 9 rules errored during the review.

[codelingo]

Avoid global variables to improve readability and reduce complexity

View Rule