Versions (relevant - OpenSearch/Dashboard/Server OS/Browser):
OpenSearch 2.11 on AWS free tier.
Describe the issue:
Hello, I have 2 deployments (different projects) on different AWS account. In both I have OpenSearch with the minimum configuration to keep the free tier per 12 months:
Availability Zone(s):
1-AZ without standby
Instance type:
t3.medium.search
Number of nodes:
1
Storage type:
EBS
EBS volume type:
General Purpose (SSD) - gp3
EBS volume size (GiB):
50
For some unexplained reason, I found that for one of the projects, after some weeks where the indexes are not queried or updated (I haven’t counted exactly how many days, but happened already twice) my indexes disappear and my login credentials are rejected, like if the node was totally cleaned and reseted.
However, when I check the snapshots, I can see the ones that contain my indexes, so I have to reindex them from there.
I haven’t added any extra configuration on policies, just the default from AWS. As I said, the only difference I noticed between the two projects is that one is being updated daily (by querying or updating items from indexes) and the other is more stable, and could or could not receive queries (less than the other, given that it’s not even public yet).
Could that be the cause? Is it normal? Is it something related to the node configuration? Does AWS automatically reset “inactive” nodes?
Thanks in advance