Skip to main content

Profiling JavaScript memory in the browser and on the server


Recently we were executing a memory usage analysis in my current (node.js) project. This article summarizes some of our findings, which may be useful to other projects. It does not describes the very basics of the mark-and-sweep garbage collection algorithm. Instead, it focuses on the details of the particular tools we used.

Chrome developer tools 

Chrome developer tools have great heap snapshot section, which can be used to analyze heaps of both browser and server (node.js) applications. It can be used to record JavaScript heap snapshots of a running process or to load existing snapshots, recorded by other tools. It is available already for several years and is extensively described on the Internet. Still, there are parts of this tool, which don’t have proper description and, overall, it may be a bit overwhelming in the beginning. So here is a very simple HTML document example, which may help in understanding Chrome heap snapshot views.

A script in this document simply creates an instance of an object, using constructor function and assigns it to global scope. If you click on “Action” and then “Set and clear” button to force browser do its initial compilation work, then record first heap snapshot, then click on “Action” button and then record second heap snapshot, you could see something similar to the picture below. Just select the second snapshot in the left menu, then select “Comparison” view and then select the first snapshot as comparison target in the dropdown, next to class filter input box.

Chrome developer tools / Profiles tab

Upper table shows the difference between two snapshots. All objects (either new or deleted in the second snapshot) are grouped by constructor function name (just as in Summary view).
First column (“Constructor”), is a tree view. On the first level you can see constructor function name (object instance grouping parameter). On the second level, you can see instances in the selected group (each instance has a unique id with “@” prefix). On the third level, you can see properties of the instances, then, on the lower level – properties of the corresponding objects and so on.
Further columns in the upper table describe difference in the number of instances and total size of each group (on the first level). On the second level, you can see whether a particular instance is new or has been deleted between the compared snapshots.

The lower (“Retainers”) table is also a tree view. On the first level, you can see list of paths from the selected instance to the root object, which causes this instance not to be marked as garbage by the garbage collector. The tree under each top level node describes the path to this root. In the picture above, you can see that instance of “MyClass” with ID “@115975” is “retained” through a property “myInst” on root object “Window”. Distance “1” means that this instance is directly referenced by a root object. It is interesting to see that there is also another, system retaining path for this instance (second top level record in the “Retainers” table). It is not directly visible to JavaScript, but it is shown in this view nonetheless.

Analyzing node.js heaps 

What is great about node.js is that not only you can build server-side using the same language (JavaScript, CoffeeScript, TypeScript, etc.) and libraries as you use for “browser-side”, but also that you can use the same tools to debug and analyze it.
There are a lot of articles, which describe how you can attach a debugger to a new or running node.js process and then either debug or record heap snapshots. What is slightly less known, is that you can actually programmatically record heap snapshots without the need to attach a debugger. That is very easy to do using “heapdump” NPM module (
For example, you can simply create a new route, which can be called any time and will create a new heap snapshot file on disk. It can even be used in production (if you closed it from unauthorized access, of course):

Using this heap dumping technique, we have very easily found that many of our processes store big unused “geoip” data files in RAM, since corresponding library loads them into memory non-lazily. Simply removing these files, greatly reduced memory consumption of our application.


Popular posts from this blog

Nginx secure link with node.js

Serving static files is a natural task for web servers. They, especially ones, having asynchronous architecture (like Nginx), are very good at such tasks.
However, usually there is an additional security logic, which should restrict access to files you've published. IIS, for example, offers deep integration with application layer, which allows custom .NET "middleware" logic injection into the request pipeline.
Node.js applications are very often published behind Nginx for various reasons and, with the help of Nginx "Secure Link" module, it's possible to offload static file serving tasks from node.js to Nginx, even if the files are not public. This module uses "shared secret" string (known to Nginx and the application) and expects a hash, based on this secret, to be present in the request to decide whether to proceed or return an error.

Secure Link module may work in 2 alternative modes (…

Performance of Redis sorted set operations

I was working on a feature recently, which involved several Redis "remove from sorted set" operations. Target production environment sorted sets are expected to be small and most of these calls would be trying to remove items, which do not exist in the sets. Although ZREM operation has documented LON(N) time complexity, ZSCORE has documented constant time complexity. This led me to believe, that Redis might have constant time complecity for ZREM calls when values to be removed do not exist in the set (if ZSCORE "knows" if item is NOT in the set in constant time, ZREM can do the same, right?).

Not in this case, apparently. ZSCORE documented constant time complexity is actually misleading (as many cases of asymptotic time complexity documentation for small data sets). Redis stores small sorted sets (up to 128 items by default) as "ziplists", which are essentially linked lists, optimized for memory consumption. Browsing through Redis source code confirms the…

Redis Lua scripts are not really transactions

Redis support of Lua scripts is a great feature. We use it a lot to build fast reliable queues with some very interesting requirements. You need it every time you want to decide your next Redis command, based on the result of a previous command, while guaranteeing that no one else has done anything with this result or anything else has changed in Redis. That is, the whole Redis script is an "atomic" operation.
However, I put it in quotes intentionally. My understanding of phrase "atomic operation" is that not only no one else can see it half complete while it is executing (that works so great in Redis). It should also mean, that it should never be left half complete if an error occurs in the middle (or at least, that is my wishful thinking:) ).
Yea, exactly, the second point doesn't work in Redis and there is no warning in the official docs. To be more polite (or precise), there is no rollback in Redis (referring to a comment in this SO question - http://stack…