listTimeSeries: increase timeout from 30 seconds to 60 seconds

The poller may often receive timeouts from the GCP monitoring API.
I have received advice from GCP support and product team to increase
the timeout of calls made from the poller cloud function.

Given that the maximum frequency of the cloud scheduler can be
once per minute, a timeout period of 60 seconds makes sense and should
reduce the amount of errors from the poller function

poller/poller-core/index.js

@@ -149,7 +149,11 @@ function getMaxMetricValue(projectId, spannerInstanceId, metric) {
     view: 'FULL'
   };
 
-  return metricsClient.listTimeSeries(request).then(metricResponses => {
+  const options = {
+    timeout: 60, // seconds
+  }
+
+  return metricsClient.listTimeSeries(request, options).then(metricResponses => {
     const resources = metricResponses[0];
     maxValue = 0.0;
     for (const resource of resources) {