mirror of
https://github.com/VictoriaMetrics/VictoriaMetrics.git
synced 2024-12-14 16:12:15 +01:00
a586b8b6d4
Previously vminsert may spread data for a single time series across all the available vmstorage nodes when vmstorage nodes couldn't handle the given ingestion rate. This could lead to increased usage of CPU and memory on every vmstorage node, since every vmstorage node had to register all the time series seen in the cluster. Now a time series may spread to maximum two vmstorage nodes under heavy load. Every time series is routed to a single vmstorage node under normal load. |
||
---|---|---|
.. | ||
insert_ctx_pool.go | ||
insert_ctx.go | ||
netstorage.go |