We are sending data to the Splunk KV store using the batch_save endpoint. Occasionally, the upload fails with this message in splunkd.log:
02-18-2016 07:39:57.706 +0000 WARN HttpListener - Read Timeout communicating with 127.0.0.1, disconnecting
I realise that the problem here is really on the client side - our process is stalling for too long in the middle of generating the data to be stored in the KV store.
However, is anyone able to confirm whether the streamInWriteTimeout setting in server.conf is the exact timeout that corresponds to the "HttpListener - Read Timeout" error message above?
streamInWriteTimeout =
* When uploading data to http server, if http server is unable to write data to
* receiver for configured streamInWriteTimeout seconds, it aborts write operation.
* Defaults to 5 second.
I've tried increasing this setting but the timeouts are so few and far-between that it's hard to know if I've applied a temporary workaround or not.
↧