Intermittent Timeouts when doing simple get

server

#1

We are experiencing’Couchbase::Error::Timeout’ and ‘Couchbase::Error::TemporaryFail’ exceptions on simple ‘find’ calls. This issue is intermittent and does not appear to effect all documents in all buckets. However, once a document returns a timeout it appears to remain inaccessible for a period of time following the first error.

I have gone through the diagnostics available to me and cannot determine an obvious source of the issue. This issue has arisen in the past two days with no obvious changes to our architecture, the views, or the data in the bucket. This bucket is linked to an ElasticSearch server via the xdcr plugin. However disabling XDCR does not appear to impact the rate at which we have this exception.

We are running CB Community 4.1.1 and using the latest Couchbase Ruby Client and Couchbase Ruby Model.


#2

After further research and experimentation. I may have misdiagnosed this.

We have multiple Views that are timing out. One of which is being accessed with stale=false. We also added some new, intensive, map-reduces which may be the culprit.

Will investigate further and share findings.


#3

Conclusion: The issue was some badly written code in our API. A loop was spamming requests to Couchbase, causing requests to timeout.