As I have not set anything referring to multiple indices or replicas (the es configuration has not been changed much at all, it's more or less as it came with the repo) I think the answer to that is no.
The node has never been part of a cluster, it's been a standalone ELK installation.
What confuses me is that it used to work this way for months, until it suddenly stopped for no apparent reason. I somehow doubt that the exhaustion of resources is the reason.
I have raised the loglevel to DEBUG and I'm getting loads of messages, most of them not meaningful to me. As far as I can see, there are no obvious or typical error messages (i. e. the words "fail", "exception" or AllocationService never appear). The word "allocation" appears frequently, typically in this context:
[2016-08-31 09:20:30,603][DEBUG][gateway ] [node-1] [filebeat-2016.07.13][2] found 1 allocations of [filebeat-2016.07.13][2], node[null], [P], v[0], s[UNASSIGNED], unassigned_info[[reason=CLUSTER_RECOVERED], at[2016-08-31T07:20:23.253Z]], highest version: [46]
[2016-08-31 09:20:30,603][DEBUG][gateway ] [node-1] [filebeat-2016.07.13][2]: throttling allocation [[filebeat-2016.07.13][2], node[null], [P], v[0], s[UNASSIGNED], unassigned_info[[reason=CLUSTER_RECOVERED], at[2016-08-31T07:20:23.253Z]]] to [[{node-1}{4AHnduKWRTOsh3Y7oXpvLg}{192.168.1.222}{192.168.1.222:9300}]] on primary allocation
Closer to the end of the logfile, lines like this appear:
2016-08-31 09:25:49,860][DEBUG][index.shard ] [node-1] [filebeat-2016.06.22][0] marking shard as inactive (inactive_time=[300000000000]) indexing wise
When I start the service, the node is in condition "red" and stays there.
shows the index building up, with an intial value of about 1090 unassigned shards, decreasing as the number of active_primary_shards and active_shards increases, until they reach 541. At that point the process stops. So far, I cannot see corresponding lines in the log (probably need to decrease the log level, DEBUG really floods you).
During startup, I am getting a lof of lines like these:
[2016-08-31 10:35:50,037][DEBUG][gateway ] [node-1] [filebeat-2016.05.25][1]: throttling allocation [[filebeat-2016.05.25][1], node[null], [P], v[0], s[UNASSIGNED], unassigned_info[[reason=CLUSTER_RECOVERED], at[2016-08-31T08:35:34.869Z]]] to [[{node-1}{1hQ6-5ZZQNO5tTmXUZRtCg}{192.168.1.222}{192.168.1.222:9300}]] on primary allocation
The last line in the log is
[2016-08-31 09:26:20,272][DEBUG][index.engine ] [node-1] [filebeat-2016.05.18][4] successfully sync committed. sync id [AVbffdFB3UQToPOebqrr].