Could anyone explain to me clearly does clustering is required or not for elastic stack?
You can most certainly run a single node Elasticsearch node without clustering, but that would naturally not be highly available and lack the added resiliency a cluster provides.
Thanks for your response, but how can I be sure, that I have to implement it, when I looking for metrics which explain what kind of situation should I begin to implement, I can't find any useful document, I know how much search and writes per second I'll get, how fast my indexes will grow, and what kind of queries my users will run, against what type of content, But I don't know at what kind of situation should I begin?
I would recommend you explain youyr use case here and provide estimates of expected data and request volumes together with what your requirements are with respect to performance, reliability and availability.That way someone can help guide you, but without any information that is virtually impossible.
You're right, I have developed a mixpanel software a few years ago for one of my clients, that has multiple software, Now a day I connect all of their software to an elastic stack to collect their user event and return to mixpannel, that admin could search over this event. Now I have collected 1000 events per second, based on my estimated this results change to 30000 events per second, and 2 or 3 admin analyze this data over the mixpannel dashboard, and As a result, I design my elastic with bool query and multi-match query, Of course at this time My elastic Stack has single nodes on the server that contain 16GB Rams and 50GB SSD. At this point should I follow up Elasticsearch clustering or not?
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.