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
Cloudflare Workers has a hard limit of 128 MB per isolate for JavaScript heap and WebAssembly allocations. When an isolate crosses that ceiling, requests can report Error 1102 (“Worker exceeded resource limits”), the dashboard shows Exceeded Memory, and analytics or Logpush records the exceededMemory outcome. The limit belongs to the isolate rather than an individual invocation, so concurrent requests can consume the same budget.
The fastest diagnosis is evidence-driven: confirm the failure signature, inspect percentile trends and deployment markers, reproduce representative traffic, compare DevTools memory snapshots for retained objects, then use a production Heap profile to find allocation call sites. Fix the specific peak or retention source—usually by streaming bodies, bounding global data, or moving durable data to KV, R2, or D1—and verify that the trend stops rising.
What does an out-of-memory failure look like in a Worker?
Start by proving that the event is a memory-limit failure rather than an ordinary JavaScript exception. Cloudflare’s documented indicators are Error 1102, the dashboard status Exceeded Memory, and the exceededMemory analytics/Logpush outcome.
Recommended Free Tools
| Where to look | Memory-limit signal | What it tells you |
|---|---|---|
| Workers dashboard | Workers & Pages → your Worker → Metrics → Errors → Invocation Statuses → Exceeded Memory | The runtime stopped an invocation because the isolate exceeded its resource limit. |
| Analytics or Logpush | exceededMemory |
A machine-readable outcome for grouping failures over time or by route. |
| Client or runtime response | Error 1102, “Worker exceeded resource limits” | The request hit a Workers resource ceiling; it is not the JavaScript-exception code. |
| Workers Logs | Filter with $metadata.error EXISTS |
Finds logged errors while you investigate a failing period. |
| Workers Logs outcome filter | Filter $workers.outcome |
Lets you inspect outcomes such as exception and separate them from memory failures. |
wrangler tail |
The exceptions field while reproducing |
Shows exception details in a live stream from the deployed Worker. |
Error 1101 is the JavaScript-exception signal, not the specific memory-limit status. A route can produce both kinds of failures at different times, so classify each event from its status or outcome before changing code.
#1 Best Overall
- Powerful Turbo Fan:CAGIWIRU Electric Air Duster have reached speed up to 130000RPM,Can efficient remove dust and debris.With LED light and Three Adjustable Speed to meet with different cleaning tasks.
- With Attached USB C Fasting Charging cable.Charing 3 hours,Enjoy up to 240 minutes of use on the lowest setting, For third speed can use for 25mins.with four charging options to meet with your needs.
- Wide Application: attached 5 different nozzles,making it suitable for a variety of scenes, such as car,pet, pc, keyboards, and other electronic devices. It also serves for office and home clean duster.
- Handy Design & Portable:Cordless air duster electric nozzle is Light weight and small size, design and comfortable to hold and space saving, convenient to carry around. You can put it in your room or inside the car without taking up space.
- Economical &Reusable: Built durable materials with a long-lasting battery,our CAGIWIRU air duster it's a sustainable alternative to disposable air cans.
What is the 128 MB limit?
Cloudflare documents 128 MB per isolate for JavaScript heap and WebAssembly allocations (Cloudflare Workers limits documentation, 2026). This is a ceiling for the running isolate, not a quota that resets for each request.
Why can one busy request affect another?
An isolate can serve multiple concurrent requests. Their live JavaScript objects, WebAssembly allocations, request data, and response data draw from the same 128 MB budget. A request that is harmless by itself can therefore fail when it overlaps with other work or with retained data from earlier requests.
How should I check whether memory is rising over time?
Look at the Workers Metrics memory chart before forming a theory about a leak. The chart uses reservoir sampling and reports P50, P90, P99, and P999 memory usage; deployment markers help connect a step change or a new slope to a release.
Read the percentiles as different operating conditions
- P50 approximates the middle of observed usage and shows the typical path.
- P90 exposes heavier request mixes that are invisible at the median.
- P99 highlights rare but operationally important peaks.
- P999 is the extreme tail and can reveal a route or payload that only occasionally approaches the ceiling.
A sustained upward trend can indicate a memory leak, but the chart cannot identify which object is being retained. Treat it as a scope-and-timing tool: note when the increase began, whether all percentiles moved, and which deployment marker is nearest.
Use deployment markers as a change-point test
- Set the chart to a period that includes a known-good deployment and the first failures.
- Compare the percentiles immediately before and after each marker.
- Record whether the change is a step (a new baseline), a gradual slope (possible retention), or isolated spikes (payload or concurrency sensitivity).
- Break the investigation down by route, method, and payload class in your logs so a noisy endpoint does not hide a smaller leak elsewhere.
Do not infer a growth rate from the chart. Cloudflare’s published material does not provide a general “normal leak rate,” and the sampled percentiles are not a substitute for object-level evidence.
How do I reproduce the problem with realistic traffic?
Reproduction is useful only when it exercises the same routes, payload shapes, and concurrency patterns that trigger production growth. Cloudflare recommends using wrangler dev, opening DevTools with D, and matching production volume and request behavior as closely as practical.
Rank #2
- Screen Cleaner is Your Screen’s New Bestfriend! - EVEO is proud to present a screen cleaner that’s perfectly suited for your TV. Meet your new screen cleaner, computer screen cleaner, laptop cleaner, iPad cleaner, Macbook,tv cleaner or any other electronic device. Our screen cleaner spray and wipe along with the included plush microfiber cloth, easily removes dust, fingerprints & other smudges on the screen’s surface. Clean effectively for a streak-free screen
- Gentle on your Screens - Our screen cleaner spray and wipes is compatible to be used even for the most sensitive LCD, LED, CRT, and OLED screens in proper use. Be at ease knowing that this product is compatible with all the major brands: LG, SONY, Samsung, Sharp, iPad, iPhone, Android screens, Kindle, PC monitors - we’ve got you covered! Can be used as LCD screen cleaner, tv cleaner, smart tv screen cleaner, and more
- Premium Super Soft Plush Microfiber Cloth - Included in this efficient screen cleaning kit. This screen cleaner spray and wipe has our signature plush microfiber cloth specially designed to comply with screens and monitors and leave them streak-free. Has been shown to effectively absorbs stubborn stains, zero streaks, and no fiber shedding. Always use this together with our screen cleaner for the best results
- Easy-to-Use Screen Cleaner Kit + microfiber cleaning cloth - The EVEO Screen Cleaning Kit will easily be your new favorite tool when it comes to taking care of the things you love most, quick and easy solution for a streak-free screen. Simply screw on the sprayer gun on the screen cleaner bottle. Spray the cleaning liquid onto the microfiber cloth, then wipe the screen until desired cleanliness & shine is achieved.
- Quality Screen Care Kit Worth Purchasing - Enjoy a convenient Screen Cleaner anytime, anywhere you need it. Your satisfaction is our top priority. We stand behind the quality of our products and that’s why, you shouldn’t be worrying at all. We manufacture the Best Screen Cleaner. and can be used for laptop cleaning kit, laptop screen cleaner, tv screen cleaner for smart tv. MacBook screen cleaner, monitor cleaner or screen cleaner spray for electronic devices with microfiber cleaning cloth
Prepare a representative test
- Warm the Worker before taking measurements; startup allocations can otherwise look like a leak.
- Include the route that fails, the largest legitimate request and response bodies, and the authentication or validation path used in production.
- Replay a sustained sequence rather than a single request so module-scope retention and concurrent overlap have time to appear.
- Keep a separate low-volume run to distinguish per-request growth from concurrency-driven peaks.
Use production-like bindings when local behavior differs
Run wrangler dev --remote when local bindings need production-like behavior. Remote mode can make KV, R2, D1, and other binding interactions more representative, while DevTools still provides the interactive inspection surface. Document the exact command, fixture sizes, request count, and concurrency for every comparison so a “fixed” run is actually comparable.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
How do DevTools memory snapshots find retained objects?
DevTools snapshots answer the liveness question: what is still reachable after the work should be finished? Capture a baseline after warm-up, exercise the suspected route, capture a second snapshot, and compare the type statistics and memory summary.
- Start the Worker with
wrangler dev. - Press D to open the Worker’s DevTools.
- Open the Memory panel and capture a snapshot after the isolate has warmed up.
- Send a representative batch of requests, including the sequence that normally precedes the alert.
- Capture a second snapshot without restarting the isolate.
- Compare Statistics and the memory summary by type; focus on types whose retained size or count increased.
- Repeat the exercise after a targeted code change and check whether the same type still grows.
Interpret a growing type as a lead, not a verdict
A larger count for strings, arrays, maps, parsed bodies, or application records tells you what to inspect next. Follow references back to the owning module, cache, closure, or request handler. Cloudflare’s example identifies a retained string caused by repeatedly appending to a module-scope variable: the snapshot reveals the live object, while the source review explains why it remains reachable.
Take snapshots at the same points in each run. A baseline taken before warm-up and a comparison taken during a different request mix can create a false difference even when the code is unchanged.
When should I use a production Heap profile?
Use a production Heap profile when you need to know where allocations originate in a deployed Worker. Cloudflare’s production profiler records a stack trace every 512 kB allocated and identifies allocating call sites. It measures allocations, not retained liveness, so it cannot by itself answer which object is still live after a request completes.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsCapture a profile from an active isolate
- Make sure the deployed Worker has been recently active and is receiving the traffic pattern that exposes the issue.
- Start a Heap profile from the dashboard, CLI, or API.
- Exercise the route and payload class that produces the high-water mark.
- Inspect hot allocation stacks and map them to parsing, validation, serialization, buffering, or application code.
- Pair the result with a DevTools snapshot when you need to distinguish temporary churn from retained objects.
A profile with a busy allocator can point to a JSON parse or schema-validation path even when those objects are promptly collected. Conversely, a small allocator that stores every result in a global array may not dominate the allocation sample but can still drive the isolate toward the ceiling. Use the two tools for their different questions rather than treating one as a replacement for the other.
Rank #3
- 【Large size】：Measuring 800 x 300 mm/ 31. 5×12 inches, which is wider and taller than others. Large mouse pad can perfectly fit your keyboard and mouse of any size, and there is also ample space for peripherals such as phones, tablets, etc.
- 【LED backlight mouse pad】：Blade Hawks RGB mouse pad provides 7 static modes and 7 dynamic modes will give you the options you are looking for. Power-off memory function remembers the last lighting mode you selected. The super glow fiber Chroma will give you a colorful gaming setup that you want.
- 【Anti-slip rubber base】：Made from Natural rubber, this gaming mousepad will firmly grip your desktop, allowing you to enjoy the ultimate precision in the games you are most passionate about. It also comes with a texture that enhances this ability even further.
- 【Smooth and waterproof surface】: Micro-textured cloth surface, for extremely precise mouse control and a smooth movement. When liquid splashes on the gaming mouse pad, it forms water droplets and slides down. Will not delay your work or gameplay.
- 【Plug and play】: This mouse and keyboard pad is powered by USB, plug and play, no driver required. One button control light modes. Press one time to change the lighting mode, press twice to change the brightness, press and hold about 3 seconds turn ON/OFF.
| Tool | Best question | Evidence produced | Main limitation |
|---|---|---|---|
| DevTools Memory snapshot | What remains live? | Object/type counts, retained size, and references at a chosen point | Interactive local reproduction; it does not show every production allocation path. |
| Production Heap profile | Where are allocations coming from? | Stack traces sampled every 512 kB allocated | Allocations are not retained memory; a hot allocator may be short-lived. |
| Metrics memory chart | When and how broadly does usage change? | P50, P90, P99, and P999 trends with deployment markers | Sampling does not identify the retaining object or code line. |
What code patterns commonly push an isolate over the limit?
Most practical fixes either lower the peak resident set or stop data from persisting across requests. Inspect the following patterns in the order suggested by your measurements.
Buffering complete request or response bodies
Reading an entire upload, download, or generated response into a string or byte array requires the complete body to be resident. If another copy is created during decoding, parsing, transformation, or serialization, the peak can be several times the wire size.
Stream bodies with TransformStream or node:stream instead of collecting the whole body first. A streaming pipeline keeps only the chunks and transformation state needed at that moment, reducing the peak that matters to the 128 MB limit. It does not make an unlimited payload safe: enforce sensible size limits and abort a stream when the request is not valid.
Unbounded module-scope arrays, strings, and caches
Module and global variables survive between requests handled by the same isolate. Appending logs, user records, response fragments, or cache entries without a bound creates retained memory that a garbage collector cannot reclaim because the global still references it.
- Replace append-only arrays with bounded queues or a fixed-size ring.
- Set explicit entry and byte limits for in-memory caches, and evict old values.
- Avoid constructing a new concatenated string on every request when the result is meant to be temporary.
- Clear per-request references in long-lived closures after the response is complete.
Keeping durable data inside the isolate
Durable or bulky state does not belong in a JavaScript global. Store it in the binding suited to the data: KV for key-value access, R2 for objects and bodies, or D1 for relational records. Read only the portion needed for a request and stream large values rather than materializing them all at once.
Schema validation dependencies
Review validation libraries when Heap profiles show allocation-heavy parsing or schema construction. Cloudflare recommends Zod 4.5.0 or later because earlier versions use substantially more memory per schema. Upgrade deliberately, rerun the same fixture set, and check both correctness and percentile memory trends; a dependency change is not proven by a single successful request.
Rank #4
- 【Softer and More Comfortable】Vaydeer wrist rest has unique diamond pattern, which is the combination of softness and aesthetics. The materials of wrist rest are improved into higher quality memory foam and covered with silky smooth lycra. The computer wrist rest makes you as comfortable and cushiony as like rest your wrists on clouds.
- 【Ergonomic Wrist Saver】The wrist rests for keyboard and mouse comes with a 17.32×3.15×0.83 inch keyboard wrist pad and a 5.94×3.15×0.83 inch mouse wrist support. Based on ergonomic design, the unique concave shape is the perfect fit for your wrist joints. The wrist rest pad fits most computer keyboards and laptops, improve hand and wrist posture, release your wrist and arm stress.
- 【Non-Slip Rubber Bottom】Featuring an anti-skid silicone base on the bottom, this wrist keyboard support stays firmly in place on your desk, preventing the padding from sliding around, ensuring stable and consistent wrist support during extended computer sessions.
- 【Better Experience & Pain Relief】Our keyboard arm rest is beneficial to alleviate the soreness caused by direct contact and friction between your arm and a hard desk surface, reducing the risk of wrist fatigue or carpal tunnel. The soft texture of memory foam can evenly distribute the pressure around your wrists and provide good support with just enough give.
- 【Helpful in Multiple Scenarios】Whether you're working, studying, writing, typing, gaming, this keyboard and mouse rest combo is an essential accessory to add comfort and support to your hands and wrists. It’s also a great gift for men, women, family, friend, coworker, gamer, teacher, etc.
What happens when the isolate reaches 128 MB?
When an isolate exceeds 128 MB, the Workers runtime lets in-flight requests complete and creates a new isolate for subsequent requests. Under extremely high load, Cloudflare says it may cancel some incoming requests to maintain platform stability.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Why a new isolate does not prove the leak is fixed
Isolate replacement can make a dashboard line fall temporarily because the new isolate starts with a smaller heap. If the code still retains data or buffers oversized bodies, the replacement isolate will grow again under the same traffic. Compare post-replacement percentiles with the pre-failure baseline and repeat the workload long enough to expose the original slope.
What “Memory limit would be exceeded before EOF” means
This runtime message indicates that buffering the remaining body would cross the memory ceiling before the stream reaches end-of-file. The complete body cannot be resident safely at that point. Streaming the body and its transformation path avoids requiring the full payload to coexist in memory, although your application must still validate size and handle backpressure.
How can I turn evidence into a targeted fix?
Use a short hypothesis-and-verification loop instead of changing several unrelated components at once.
- State the failure precisely. Record the route, invocation status, payload class, concurrency, and whether the event is Error 1102,
exceededMemory, or Error 1101. - Choose the evidence that matches the question. Use Metrics for timing and scope, snapshots for retained objects, and a production Heap profile for allocating call sites.
- Form one bounded hypothesis. Examples include “the response body is copied twice,” “a module-scope array grows per request,” or “schema construction allocates heavily.”
- Make the smallest relevant change. Stream the body, bound or remove the global, move state to KV/R2/D1, or update the validation dependency.
- Replay the same workload. Keep route mix, fixture sizes, warm-up, and concurrency consistent with the baseline.
- Verify all three outcomes. Confirm that the relevant object no longer grows in snapshots, the allocation stack changes as expected in a Heap profile, and percentile memory trends stop rising.
- Watch the deployed release. Use deployment markers and invocation outcomes after rollout; a local improvement is not enough if production traffic has a different concurrency or payload distribution.
Separate peak reduction from persistence fixes
Streaming is primarily a peak-memory fix: it lowers how much of a body must be resident simultaneously. Moving data to KV, R2, or D1 is primarily a persistence fix: it removes durable state from the isolate. A Worker can need both changes when a retained cache and a large buffered body contribute to the same failure.
Which diagnostic path fits each symptom?
| Observed symptom | First check | Likely next action |
|---|---|---|
| All percentiles jump immediately after a release | Deployment marker and changed module initialization | Take a warm baseline snapshot and inspect newly created globals, caches, or preloaded data. |
| P50 is stable but P99/P999 spike on large requests | Route and payload breakdown | Stream request/response bodies, cap payload size, and inspect temporary copies in parsing or serialization. |
| Memory rises after every batch of requests | Two snapshots around the same workload | Trace growing arrays, strings, maps, closures, and cache entries back to module or request scope. |
| Heap profile shows one hot allocator but snapshots remain flat | Allocation stack and garbage-collection behavior | Reduce allocation churn if it causes peaks, but do not label it a leak without retained-object growth. |
| Production fails but local reproduction does not | Traffic fidelity and binding mode | Use representative volume and payloads; try wrangler dev --remote for production-like bindings, then profile an active deployment. |
| “Memory limit would be exceeded before EOF” | Body-handling code path | Replace whole-body buffering with a streaming pipeline and enforce an explicit size policy. |
What should I record during an investigation?
A concise incident record prevents misleading comparisons between runs. Capture:
Best Value
- 【No Drill Mounted】Cable management tray can be assembled in as little as 3 minutes or less-simply. Side clips easily clampe the edge of desk, pass through the holes on the sides, and secure excess cables with baskets. Desk cable management supports inward or outward installation to meet your needs in different scenarios, and provides you with great convenience when collecting and organizing wires.
- 【Rubber Pad Protects Desk】Wire organizer under desk built-in soft rubber pads on the clamp protect your desk from scratches, so you will not damage your furniture or office.
- 【Sturdy and Beautiful】Cord management under desk is made of high-quality carbon steel and holds about 15 lbs. The wire basket has 2 holes on each side of the tray to enable your cables to pass freely. Cable baskets are suitable for desk thickness use from 0.4" to 2.4".
- 【Space Saving & Keep Organized】Desk cable management tray holds all your power cords, outlet strips, USB hubs and computer transformers hidden from your feet, and the data cables will not enter the vacuum cleaner to keep your desk clean and tidy. Desk wire organizer is also suitable for use in the kitchen and outdoors.
- 【Reduce Safety Risks】Cable tray under desk keeps all plugs away from your kids, no more worry about tripping. Using 2 holes wire tray saves extra space, helps you easily cabling and wire protection. The mesh design is not easy to accumulate dust, but avoids safety accidents caused by messy wires.
- Worker name, deployment identifier, route, method, and region or traffic context used for the test.
- Invocation status and outcome fields, including whether the event was Error 1102,
exceededMemory, or Error 1101. - Metrics time range, percentile values, and the nearest deployment marker.
- Warm-up duration, request count, concurrency, fixture sizes, and whether
--remotewas used. - Snapshot timing and the object types whose retained size changed.
- Heap-profile timing and the top allocation stacks sampled at 512 kB intervals.
- The exact code or dependency change and the post-change verification window.
This record lets you distinguish a real reduction from a quieter test, a fresh isolate, or a workload that never reached the failing path.
What is the practical checklist for preventing a recurrence?
- Keep the 128 MB per-isolate ceiling visible in design reviews for large payload and WebAssembly features.
- Prefer streaming APIs for bodies that can exceed a small, known bound.
- Never allow module-scope collections or caches to grow without explicit count and byte limits.
- Store durable and bulky data in KV, R2, or D1 rather than in globals.
- Track memory percentiles and deployment markers after releases that change parsing, validation, caching, or serialization.
- Use snapshots to prove retention and production Heap profiles to locate allocation sources.
- Test with concurrent, production-shaped traffic instead of a single happy-path request.
- Keep validation dependencies current; Cloudflare specifically recommends Zod 4.5.0 or later for lower per-schema memory use.
- Alert on
Exceeded MemoryandexceededMemoryseparately from JavaScript exceptions so the response matches the failure.
Frequently Asked Questions
Can a paid Workers plan increase the 128 MB memory limit?
No. Cloudflare’s limits documentation keeps memory at 128 MB per isolate; plan differences that increase CPU limits do not raise this memory ceiling. Treat plan selection and memory remediation as separate decisions.
Why does a local memory snapshot look healthy while production still fails?
Local tests often use fewer concurrent requests, smaller payloads, or different binding behavior. Reproduce the production route mix and volume, use remote bindings when appropriate, and capture a profile from a recently active deployed isolate.
Free tools Windows power users keep installed
One-click scans. No signup required.
Is a high allocation rate automatically a memory leak?
No. A Heap profile samples allocation call sites, while a leak requires objects to remain reachable. Confirm retention with before-and-after snapshots; temporary allocation churn may need optimization for peak usage but is not proof of a leak.
Will restarting or replacing an isolate permanently clear the problem?
It clears the current heap and can make metrics drop briefly, but the same code will grow again if it retains data or buffers oversized bodies. Use the replacement as an observation point, then verify the workload after the fix.
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.
