Skip to main content

Bandwidth and Concurrency Tuning

Backup jobs should be predictable. Tune them so they finish on time without hurting production traffic.

Control Flags

FlagPurpose
--bwlimitLimit transfer rate
--transfersConcurrent file upload/download workers
--checkersMetadata compare workers
--tpslimitAPI transactions per second cap

Mapping

Example Profiles

office-hours-profile.sh
rclone sync /srv/data remote-prod:backup/data \
--bwlimit 20M \
--transfers 6 \
--checkers 12 \
--tpslimit 8
nightly-profile.sh
rclone sync /srv/data remote-prod:backup/data \
--bwlimit 0 \
--transfers 16 \
--checkers 32

Scheduling Strategy

Time windowProfile
Business hoursConservative bandwidth and TPS
Off-peak/nightHigher throughput profile
tip

Store profiles in scripts so operators can switch modes safely without hand-editing command lines.

Common Pitfalls

PitfallImpactMitigation
Unlimited daytime bandwidthApp latency spikesSet --bwlimit
High --transfers with low CPUQueueing and retriesMatch concurrency to host resources
No API TPS capProvider throttle burstsAdd --tpslimit

What's Next