Cant seem to get past the connecting Kibana to Elasticsearch when starting Snowplow mini 0.6.0
When navigating to Kibana the following error is displayed:
plugin:elasticsearch@6.3.1 Unable to connect to Elasticsearch at http://elasticsearch:9200.
I’ve double checked the elasticsearch.yml configuration and network.host: 0.0.0.0 is set accordingly.
Hi @ben ,
That error is normal if you see that just after the instance launch, it may take Kibana ~10 sec to connect to Elasticsearch, please refresh the page after few seconds to check if error still stands.
If error stands, there might be an issue with service launch in which case I’d recommend to restart services through control plane, wait few seconds and hit /kibana endpoint again.
Let us know how troubleshooting goes.
Hi @ben,
What instance type and AMI have you selected? It could be that you have selected an instance without sufficient memory which is preventing Elasticsearch from launching successfully.
I am seeing the same error. I logged into Snowplow-mini’s ec2 instance and saw that the docker container for elasticsearch is forever in “Restarting” state.
Here’s the instance type I’m using: snowplow-mini-0.6.0-large-1533804941-hvm-ebs-amd64 (ami-09e9b4ba16b4682a4) on t2.medium
@rahulj51 - after poking around a bit more, I noticed the same behavior of the elasticsearch container constantly trying to restart itself.
@josh - that’s also interesting… I feel I saw a message in one of the logs regarding insufficient memory. I’ll restart my instance again shortly and confirm.
Hey @ben the latest release requires a “large” instance not a “medium” instance. The memory requirements are hardcoded into the AMI and require more memory than is available on the “medium”.
To get this working you will need to use a t2.large or m5.large node type (or equivalent specification).