iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Storage slows under load when the workload’s requests reach a limit somewhere along the I/O path—or when too few requests are in flight to use the available capacity. To find the cause, read IOPS, throughput, latency, and queue depth together, then compare them with both the storage volume’s limits and the host or instance’s limits.
What changes when storage is under load?
An application issues read or write requests. The operating system and device path may queue them; the storage system processes them; and the application waits for acknowledgement. When demand stays within the slowest part of this path, requests complete at a steady rate. When demand exceeds that capacity, requests wait and latency can rise. If the application has too little concurrent work, a capable volume can instead remain underused.
That makes “disk utilization” or a single performance number an incomplete diagnosis. The workload’s request size and access pattern, the number of outstanding requests, and every limit between application and storage shape the result.
What do IOPS, throughput, latency, and queue depth mean?
- IOPS counts completed input/output operations per second. It describes operation rate, not how much data each operation moves.
- Throughput measures the amount of data transferred over time, commonly expressed in MiB/s.
- Latency is the time an I/O takes from the client’s request to acknowledgement. In AWS EBS terminology, it is end-to-end client time.
- Queue depth is the number of I/O requests waiting or in flight. AWS defines its volume queue length as the number of pending I/O requests for a device.
Request size connects IOPS and throughput: roughly, data rate depends on operations per second multiplied by bytes per operation. Many small requests can produce high IOPS but modest throughput. Large requests move more data per operation, so a throughput ceiling can be reached at a lower IOPS rate.
#1 Best Overall
- Easily store and access 2TB to content on the go with the Seagate Portable Drive, a USB external hard drive
- Designed to work with Windows or Mac computers, this external hard drive makes backup a snap just drag and drop
- To get set up, connect the portable hard drive to a computer for automatic recognition no software required
- This USB drive provides plug and play simplicity with the included 18 inch USB 3.0 cable
- The available storage capacity may vary.
Why can a high IOPS limit still deliver modest throughput?
A specific AWS EBS example shows why the two limits must be considered separately. AWS documentation describes a gp2 volume under 1,000 GiB, with burst credits available, as having an IOPS limit of 3,000 and a volume throughput limit of 250 MiB/s. With 256 KiB requests, the throughput limit is reached at 1,000 IOPS: 1,000 × 256 KiB equals 250 MiB/s. These figures apply to that documented AWS configuration, not to storage in general.
When diagnosing a real workload, use its actual request size rather than assuming every I/O is small or large. A workload can be below its IOPS ceiling yet already constrained by throughput.
Rank #2
- Easily store and access 5TB of content on the go with the Seagate portable drive, a USB external hard Drive
- Designed to work with Windows or Mac computers, this external hard drive makes backup a snap just drag and drop
- To get set up, connect the portable hard drive to a computer for automatic recognition software required
- This USB drive provides plug and play simplicity with the included 18 inch USB 3.0 cable
- The available storage capacity may vary.
How should queue depth and latency be read together?
A rising queue alongside high latency is consistent with requests arriving faster than the system can complete them. It does not, by itself, prove the volume is the bottleneck: the limiting point could also be the instance, network link, guest operating system, or another part of the path.
Recommended Free Tools
A low queue is not automatically evidence of healthy performance or spare capacity. If latency and utilization are low while the application is slow, it may not be issuing enough concurrent I/O to use the capacity available. Conversely, increasing concurrency without regard to request size and latency can create a bottleneck elsewhere.
Rank #3
- Easily store and access 1TB to content on the go with the Seagate Portable Drive, a USB external hard drive.Specific uses: Personal
- Designed to work with Windows or Mac computers, this external hard drive makes backup a snap just drag and drop. Reformatting may be required for Mac
- To get set up, connect the portable hard drive to a computer for automatic recognition no software required
- This USB drive provides plug and play simplicity with the included 18 inch USB 3.0 cable
- The available storage capacity may vary.
AWS offers a specific EBS guideline: for maximum consistency, a volume should maintain an average queue depth, rounded to the nearest whole number, of one for every 1,000 provisioned IOPS in a minute. Treat this as AWS guidance for EBS, not a universal target for other storage systems or every workload. Queue depth should be calibrated with I/O size and latency.
How does workload shape change what good performance looks like?
There is no universal storage ranking that predicts performance under every workload. The relevant combination is access pattern, request size, response-time sensitivity, product configuration, and limits across the full I/O path. AWS’s descriptions below apply to Amazon EBS volume families specifically.
Rank #4
- Easily store and access 4TB of content on the go with the Seagate Portable Drive, a USB external hard drive.Specific uses: Personal
- Designed to work with Windows or Mac computers, this external hard drive makes backup a snap just drag and drop
- To get set up, connect the portable hard drive to a computer for automatic recognition no software required
- This USB drive provides plug and play simplicity with the included 18 inch USB 3.0 cable
- The available storage capacity may vary.
| Workload or access pattern | What matters | AWS EBS guidance |
|---|---|---|
| Random or sequential I/O | Request rate and data rate depend on the workload and volume configuration. | AWS describes SSD-backed EBS volumes as delivering consistent performance for either pattern. |
| Large sequential I/O | Throughput may matter more than low response time. | AWS describes the st1 and sc1 HDD-backed families as optimized for large sequential I/O. |
| Transaction-intensive work | Increased I/O latency can affect response time. | AWS says these workloads are sensitive to increased latency and well-suited to SSD-backed volumes. |
| Throughput-intensive work | Bulk data rate may take priority over latency. | AWS says these workloads are less sensitive to increased latency and well-suited to HDD-backed volumes. |
These are workload tendencies, not immutable rules. For example, a throughput-oriented HDD volume can be a poor match for excessive small random I/O; check the product’s documented limits against the actual request pattern.
Free tools Windows power users keep installed
One-click scans. No signup required.
Can the host or instance be the bottleneck instead of the volume?
Yes. For Amazon EBS, performance depends on both the volume’s capability and sufficient bandwidth at the attached instance. An instance’s aggregate EBS bandwidth can constrain traffic even when the volume itself has headroom. Compare measured IOPS and throughput with both sets of limits before concluding that the volume is the sole problem. AWS’s EBS optimization documentation discusses the instance-side context; its limits and recommendations are AWS-specific.
Best Value
- [Upgraded Version] - This external hard drive features a mirrored logo stripe combined with a striped anti-slip design, and the rounded corners of the casing make it easier to grip. The stripes also have a heat dissipation function, ensuring stable and fast data transfer.
- 【Ultra-thin and quiet】 - The motherboard adopts JMicron 578 noise-free solution, giving you a quiet working environment. Lightweight and portable size designed to fit in your pocket for easy portability.
- 【Ultra-Fast Data Transfers】 - Pairing this external hard drive with JMicron 578 solution USB 3.0 and USB 2.0 interfaces enables blazing-fast data transfer. It boasts theoretical read speeds of up to 125MB/s and write speeds of up to 103MB/s.
- 【Plug and Play】 - With no software to install, just plug it in and the drive is ready to use.The hard disk chip is wrapped with an aluminum anti-interference layer to increase heat dissipation and protect data.
- 【What You Get】 - 1 x Portable Hard Drive, 1 x USB 3.0 Cable, 1 x User Manual, Gift-type shell packaging ,Three-year manufacturer's warranty and free technical support services.
AWS also states that io2 Block Express volumes are designed to deliver average latency under 500 microseconds for 16 KiB I/O operations in the stated EBS-optimized-instance context. That is an AWS design target with those conditions, not a general storage benchmark or a promise for every workload.
How to diagnose slow storage under load
- Define the symptom and interval. Establish whether reads, writes, or both are affected; identify typical request size and random versus sequential behavior; and determine whether the slowdown is sustained or occurs in short bursts.
- Compare demand with both capacity limits. Check observed IOPS and throughput against the configured volume limits and the attached instance’s EBS limits. If the workload is reaching a limit, review whether the capacity or volume configuration fits the workload.
- Inspect latency and queue depth together. A growing queue and high latency suggest completion is falling behind demand. A low queue with low utilization may point to insufficient application concurrency rather than inadequate provisioned capacity.
- Check for path and workload-specific conditions. Look for an instance bandwidth ceiling, first access to blocks on a snapshot-created volume, or a mismatch such as heavy small random I/O on a throughput-oriented HDD volume.
- Use metrics at a useful time scale. Review average read and write latency, IOPS, throughput, and exceeded-limit checks. Minute averages can hide short bursts, so use detailed per-I/O statistics or histograms when supported by the instance and platform.
Which storage metrics should you monitor?
For EBS, AWS recommends tracking average read/write latency, IOPS, throughput, and exceeded-limit checks. Interpret the metrics over the interval that matches the symptom: an average over a minute may smooth away a brief but disruptive burst. AWS documents detailed real-time performance statistics for supported Nitro-based instances; availability depends on the instance and platform.
Snapshot-created volumes have an additional consideration: first access to each block may have higher latency. If a slowdown appears during initial reads after volume creation, account for this behavior before treating it as steady-state performance.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
