You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The initial implementation was done with strict (as much as possible) consistency guarantees. However, with ~25 telemetry sets per second (from 525 NEs with 30-sec intervals) and with the subscription on network-rx/tx, disk, load, on/off -- the throughput of the implementation wasn't able to catch up and was starting to create a backlog.
Solution: trade the consistency for throughput and implement local cache and bulk ES operations. Downside: in case of backlog (not due to the service), the multiple instances of the service may create a race condition.
The text was updated successfully, but these errors were encountered:
The initial implementation was done with strict (as much as possible) consistency guarantees. However, with ~25 telemetry sets per second (from 525 NEs with 30-sec intervals) and with the subscription on network-rx/tx, disk, load, on/off -- the throughput of the implementation wasn't able to catch up and was starting to create a backlog.
Solution: trade the consistency for throughput and implement local cache and bulk ES operations. Downside: in case of backlog (not due to the service), the multiple instances of the service may create a race condition.
The text was updated successfully, but these errors were encountered: