Hi,
I am running an ES (version 6.3.1) cluster on AWS's managed-ES service, and have a script (written in Python) that is running every so often to check for index sizes, and if any index exceeds a given threshold, I will set that index's index.blocks.write setting to true via 'curl --fail --request PUT "%(_host)s/%(index)s/_settings" -H \'Content-Type: application/json\' -d "{ \\"index\\": { \\"blocks.write\\": true } }"'.
Given every time I make the curl request above, the answer is always {"acknowledged":true}, I am having hard time understanding, and troubleshooting, why clients are still able to write to now, in theory, write-blocked indices via the _bulk endpoint anyway.
Is there any setting that allows clients to set back indices to write-allowed mode automagically? Is there anyway I can see what and reason an index goes back to blocks.write: false "by itself"?
Thank you.