Shard size does affect query latency, so if your CPU is not saturated during querying, you may benefit from decreasing the shard size, e.g. by increasing the number of primary shards to 4 or 6 instead of 3. To benefit from indices covering a smaller time period I do however suspect you need to migrate to 5.x.
kk, it helps. thank you
still I just did a odd test ( for the fun ) . tried to fetch "only in discover" the last 30 days, and it work . no timeout but takes long to load..
so the problems is really upon visualisation ...
Discover just fetches a sample of 500 documents and do not aggregate across all records, so has little in common with the aggregations behind visualizations.
So long story made short... I would have to change the amount of shards to 6 and it will manage the "spread" by it self ?
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.