Skip to main content

Linux network rate limit

Last updated on

Linux network rate limit is a chaos fault that throttles egress bandwidth on NETWORK_INTERFACES of the target Linux machine to NETWORK_BANDWIDTH (with a burst of BURST and queue limit LIMIT) for DURATION, then restores normal connectivity. Rate limiting is restricted to traffic destined for DESTINATION_HOSTS/DESTINATION_IPS and the configured port filters; SSH ports stay reachable when WHITELIST_SSH is true. The fault runs through the Linux Chaos Infrastructure (LCI) systemd service installed on the target VM.

Use this fault to test how a workload behaves when bandwidth is constrained: whether large transfers degrade gracefully, whether back-pressure flows through the application correctly, whether buffer pools and queue depths stay inside bounds, and whether monitoring detects the saturation within the alerting SLA.

Run your first experiment

If you have not installed the Linux Chaos Infrastructure yet, go to Linux Chaos Infrastructure to install the agent and connect the VM to the control plane.

Use cases​

Run this fault when you want to answer concrete questions like:

  • Bandwidth headroom: When egress is throttled to NETWORK_BANDWIDTH, do bulk transfers complete inside the SLA?
  • Back-pressure: Do producers slow down cleanly when the pipe is full, or do they OOM the local buffer?
  • Queue depth: Do queue depths in the application stay inside bounds when downstream is throttled?
  • Monitoring fidelity: Do alerts on transmit queue length, congestion, and SLA breach fire within the alerting SLA?

Prerequisites​

  • Linux Chaos Infrastructure installed: The linux-chaos-infrastructure systemd service is active on the target VM and the infrastructure is in CONNECTED state. Go to Linux Chaos Infrastructure to install it.
  • Target interface exists: Each entry in NETWORK_INTERFACES exists on the target VM. Confirm with ip -br link.
  • tc available: The fault uses Linux Traffic Control (tc) with the token bucket filter (tbf) qdisc. Provided by iproute2.

Supported environments​

The fault has been tested on the following Linux distributions. Go to Linux fault requirements to see the full compatibility matrix.

PlatformSupport status
Ubuntu 16+, Debian 10+Supported
CentOS 7+, RHEL 7+, Fedora 30+Supported
openSUSE LEAP 15.4+ / SUSE Linux Enterprise 15+Supported

Permissions required​

This fault is classified as an Advanced Linux fault. It requires the Linux Chaos Infrastructure systemd service to run with the root user and root user group on the target VM so it can manage the tc qdisc. No cloud credentials are needed.

Fault tunables​

Configure the following fault parameters when you add Linux network rate limit to an experiment in Chaos Studio. Defaults are shown for reference.

Chaos parameters

TunableDescriptionDefault
DURATIONTotal duration of the fault. Accepts [hours]h[minutes]m[seconds]s format.30s
NETWORK_BANDWIDTHBandwidth ceiling for egress traffic (for example, 1mbit, 512kbit).1mbit
BURSTBurst size that the bucket allows above NETWORK_BANDWIDTH (for example, 32kb).32kb
LIMITQueue limit before packets start to be dropped (for example, 2mb).2mb
PEAK_RATEPeak rate ceiling (for example, 2mbit). Leave empty to skip the peak rate cap.""
MIN_BURSTMinimum burst size, in bytes, used with PEAK_RATE. Leave empty to skip.""
NETWORK_INTERFACESComma-separated network interfaces to apply chaos on.eth0
RAMP_TIMEWait period in seconds before and after the fault. Go to ramp time to read how it is applied.0

Target filters (provide at least one host or IP to limit blast radius)

TunableDescriptionDefault
DESTINATION_HOSTSComma-separated destination host names to target.""
DESTINATION_IPSComma-separated destination IPs to target. Per-IP ports can be specified using the ip|port format.""
SOURCE_PORTSComma-separated source ports to target. Prefix with ! to whitelist.""
DESTINATION_PORTSComma-separated destination ports to target. Prefix with ! to whitelist.""
WHITELIST_SSHKeep SSH ports (22,2222) excluded from chaos.true

When neither DESTINATION_HOSTS nor DESTINATION_IPS is set, the fault applies to all destinations on the interface.

Tunables that apply to every fault are documented in common tunables for all faults.

Fault execution in brief​

Adds a token bucket filter (tbf) qdisc on NETWORK_INTERFACES that caps egress bandwidth at NETWORK_BANDWIDTH (with BURST and LIMIT) for DURATION against the configured filters, then removes the qdisc.

Expected behavior during fault execution​

  • Egress throughput to matched destinations is capped at approximately NETWORK_BANDWIDTH.
  • Burst traffic up to BURST is allowed; sustained traffic is throttled.
  • The transmit queue grows until it hits LIMIT, after which packets are dropped.
  • Application bulk transfers slow down; small request/response patterns may still complete inside the SLA.
  • After the duration ends, the qdisc is removed and bandwidth returns to baseline.
When the fault ends

The chaos pod removes the tbf qdisc. Bandwidth returns to baseline; queued packets drain into the underlying interface.

Signals to watch​

Attach resilience probes to assert each layer:

  • Egress throughput: Use a Prometheus probe on rate(node_network_transmit_bytes_total[1m]) and assert it dropped to the configured ceiling.
  • Queue depth: Use a command probe running tc -s qdisc show dev <interface> and inspect the qdisc stats.
  • End-to-end availability: Use an HTTP probe on a user-visible bulk endpoint.

Verify the fault execution effect​

  1. Inspect the active qdisc.

    tc -s qdisc show dev <interface>

    A tbf qdisc with the configured rate and burst should be present during the chaos window.

  2. Run a bandwidth test against a matched destination.

    iperf3 -c <target> -t 30

    Throughput should cap at approximately NETWORK_BANDWIDTH during the chaos window and return to line rate afterwards.

  3. Inspect Linux Chaos Infrastructure logs.

    sudo journalctl -u linux-chaos-infrastructure -n 100 --no-pager

Recovery and cleanup​

  • End of duration: The chaos pod removes the tbf qdisc when DURATION elapses.
  • Abort the experiment: Stopping the experiment from Chaos Studio also removes the qdisc.
  • Manual recovery: If the qdisc survives an abort, remove it with sudo tc qdisc del dev <interface> root.

Limitations​

  • Egress only: The fault throttles packets leaving the VM on the configured interfaces; ingress bandwidth is not capped.
  • Single VM scope: Each fault run targets one VM.
  • TBF behavior: Token bucket filtering allows short bursts above NETWORK_BANDWIDTH; tune BURST to constrain bursting.
  • SSH whitelisting: When WHITELIST_SSH is false, your SSH session may slow down or stall if SSH ports match the filter.

Troubleshooting​

Linux network rate limit fault shows no measurable throttling in Harness Chaos Engineering

Confirm the destination filter matches your test traffic. Run tc -s qdisc show dev <interface> to verify the tbf qdisc is installed. NETWORK_BANDWIDTH values above the line rate have no effect.

Bursts above NETWORK_BANDWIDTH visible in monitoring

Token bucket filtering allows short bursts up to BURST. Reduce BURST (for example, to 1kb) to constrain bursting. PEAK_RATE caps the maximum instantaneous rate when set.

Throttling persists after the experiment ends

If the tbf qdisc was not removed, remove it manually with sudo tc qdisc del dev <interface> root.