Multiple MONSTACHE_ES_URLS not used : When first node is down, monstache doesn't seem to be trying to connect to another node

[iandeveseleer]

Hello,

We are using Monstache v6.7.7 on a ElasticSearch v7.17.26 cluster with 5 nodes. Recently one of the 5 nodes of our ElasticSearch cluster, the first, was down for a day, and we then noticed that the system doesn't manage the list of elasticsearch nodes urls provided, and never attempts to connect to another node that's still accessible.

Our monstache is running in a Kubernetes statefulset, where a Docker image is used with an entrypoint starting the monstache, with MONSTACHE_ES_URLS valued with the list of nodes urls before.

The down state of our node-001 has led our statefulset to fall in crashloopbackoff state.

Here is the logs of Monstache when trying to connect to node-001, where we seen there is no try to use node-[002-005] urls :

INFO 2025/02/11 18:57:53 Started monstache version 6.7.7
INFO 2025/02/11 18:57:53 Go version go1.14.10
INFO 2025/02/11 18:57:53 MongoDB go driver v1.8.0
INFO 2025/02/11 18:57:53 Elasticsearch go driver 7.0.28
INFO 2025/02/11 18:57:53 Successfully connected to MongoDB version 7.0.16
ERROR 2025/02/11 18:57:54 Unable to validate connection to Elasticsearch: Get "http://node-001:9200/": dial tcp X.X.X.X:9200: connect: connection refused

The MONSTACHE_ES_URLS is well valued, with "," separator between each node url.

Thank you in advance for your assistance !

[ginolegigot]

Hello!
Same issue here on a similar config