Fix flaky test

[delvedor]

The following test passes consistently in localhost, but it's flaky on CI. It should be refactored to avoid this issue. Very likely the problem is caused due to the use of timers.

elasticsearch-js/test/behavior/sniff.test.js

Lines 114 to 149 in 94d4ac7

	test('Sniff interval', { skip: 'Flaky on CI' }, t => {
	t.plan(10)
	buildCluster(({ nodes, shutdown, kill }) => {
	const client = new Client({
	node: nodes[Object.keys(nodes)[0]].url,
	sniffInterval: 50
	})
	// this event will be triggered by api calls
	client.on(events.SNIFF, (err, request) => {
	t.error(err)
	const { hosts, reason } = request.meta.sniff
	t.strictEqual(
	client.connectionPool.size,
	hosts.length
	)
	t.strictEqual(reason, Transport.sniffReasons.SNIFF_INTERVAL)
	})
	t.strictEqual(client.connectionPool.size, 1)
	setTimeout(() => client.info(t.error), 60)
	setTimeout(() => {
	// let's kill a node
	kill('node1')
	client.info(t.error)
	}, 150)
	setTimeout(() => {
	t.strictEqual(client.connectionPool.size, 3)
	}, 200)
	t.teardown(shutdown)
	})
	})

[penge]

Do you have the log from the CI?

Wasn't the reason for the fail TimeoutError by any chance?

[delvedor]

This is the log for the failure: https://github.com/elastic/elasticsearch-js/runs/489895557.

You can reproduce it yourself by forking this repository and playing with GitHub actions :)

[penge]

Cool! :) I'll check it out!