Category: Azure Functions

Azure Functions vs AWS Lambda vs Google Cloud Functions – JavaScript Scaling Face Off

Since I published this piece Microsoft have made significant improvements to HTTP scaling on Azure Functions and the below is out of date. Please see this post for a revised comparison.

I had a lot of interesting conversations and feedback following my recent post on scaling a serverless .NET application with Azure Functions and AWS Lambda. A common request was to also include Google Cloud Functions and a common comment was that the runtimes were not the same: .NET Core on AWS Lambda and .NET 4.6 on Azure Functions. In regard to the latter point I certainly agree this is not ideal but continue to contend that as these are your options for .NET and are fully supported and stated as scalable serverless runtimes by each vendor its worth understanding and comparing these platforms as that is your choice as a .NET developer. I’m also fairly sure that although the different runtimes might make a difference to outright raw response time, and therefore throughput and the ultimate amount of resource required, the scaling issues with Azure had less to do with the runtime and more to do with the surrounding serverless implementation.

Do I think a .NET Core function in a well architected serverless host will outperform a .NET Framework based function in a well architected serverless host? Yes. Do I think .NET Framework is the root cause of the scaling issues on Azure? No. In my view AWS Lambda currently has a superior way of managing HTTP triggered functions when compared to Azure and Azure is hampered by a model based around App Service plans.

Taking all that on board and wanting to better evidence or refute my belief that the scaling issues are more host than framework related I’ve rewritten the test subject as a tiny Node / JavaScript application and retested the platforms on this runtime – Node is supported by all three platforms and all three platforms are currently running Node JS 6.x.

My primary test continues to be a mixed light workload of CPU and IO (load three blobs from the vendors storage offering and then compile and run a handlebars template), the kind of workload its fairly typical to find in a HTTP function / public facing API. However I’ve also run some tests against “stock” functions – the vendor samples that simply return strings. Finally I’ve also included some percentile based data which I obtained using Apache Benchmark and I’ve covered off cold start scenarios.

I’ve also managed to normalise the axes this time round for a clearer comparison and the code and data can all be found on GitHub:

(In the last week AWS have also added full support for .NET Core 2.0 on Lambda – expect some data on that soon)

Gradual Ramp Up

This test case starts with 1 user and adds 2 users per second up to a maximum of 500 concurrent users to demonstrate a slow and steady increase in load.

The AWS and Azure results for JavaScript are very similar to those seen for .NET with Azure again struggling with response times and never really competing with AWS when under load. Both AWS and Azure exhibit faster response times when using JavaScript than .NET.

Google Cloud Functions run fairly close to AWS Lambda but can’t quite match it for response time and fall behinds on overall throughput where it sits closer to Azure’s results. Given the difference in response time this would suggest Azure is processing more concurrent incoming requests than Google allowing it to have a similar throughput after the dip Azure encounters at around the 2:30 mark – presumably Azure allocates more resource at that point. That dip deserves further attention and is something I will come back to in a future post.

Rapid Ramp Up

This test case starts with 10 users and adds 10 users every 2 seconds up to a maximum of 1000 concurrent users to demonstrate a more rapid increase in load and a higher peak concurrency.

Again AWS handles the increase in load very smoothly maintaining a low response time throughout and is the clear leader.

Azure struggles to keep up with this rate of request increase. Response times hover around the 1.5 second mark throughout the growth stage and gradually decrease towards something acceptable over the next 3 minutes. Throughput continues to climb over the full duration of the test run matching and perhaps slightly exceeding Google by the end but still some way behind Amazon.

Google has two quite distinctively sharp drops in response time early on in the growth stageas the load increases before quickly stabilising with a response time around 140ms and levels off with throughput in line with the demand at the end of the growth phase.

I didn’t run this test with .NET, instead hitting the systems with an immediate 1000 users, but nevertheless the results are inline with that test particularly once the growth phase is over.

Immediate High Demand

This test case starts immediately with 400 concurrent users and stays at that level of load for 5 minutes demonstrating the response to a sudden spike in demand.

Both AWS and Google scale quickly to deal with the sudden demand both hitting a steady and low response time around the 1 minute mark but AWS is a clear leader in throughput – it is able to get through many more requests per second than Google due to its lower response time.

Azure again brings up the rear – it takes nearly 2 minutes to reach a steady response time that is markedly higher than both Google and AWS. Throughput continues to increase to the end of the test where it eventually peaks slightly ahead of Google but still some way behind AWS. It then experiences a fall off which is difficult to explain from the data available.

Stock Functions

This test uses the stock “return a string” function provided by each platform (I’ve captured the code in GitHub for reference) with the immediate high demand scenario: 400 concurrent users for 5 minutes.

With the functions essentially doing no work and no IO the response times are, as you would expect, smaller across the board but the scaling patterns are essentially unchanged from the workload function under the same load. AWS and Google respond quickly while Azure ramps up more slowly over time.

Percentile Performance

I was unable to obtain this data from VSTS and so resorted to running Apache Benchmarker. For this test I used settings of 100 concurrent requests for a total of 10000 requests, collected the raw data, and processed it in Excel. It should be noted that the network conditions were less predictable for these tests and I wasn’t always as geographically close to the cloud function as I was in other tests though repeated runs yielded similar patterns:

AWS maintains a pretty steady response time up to and including the 98th percentile but then shows marked dips in performance in the 99th and 100th percentiles with a worst case of around 8.5 seconds.

Google dips in performance after the 97th percentile with it’s 99th percentile roughly equivalent to AWSs 100th percentile and it’s own 100th percentile being twice as slow.

Azure exhibits a significant dip in performance at the 96th percentile with a sudden drop in response time from a not great 2.5 seconds to 14.5 seconds – in AWSs 100th percentile territory. Beyond the 96th percentile their is a fairly steady decrease in performance of around 2.5 seconds per percentile.

Cold Starts

All the vendors solutions go “cold” after a time leading to a delay when they start. To get a sense for this I left each vendor idle overnight and then had 1 user make repeat requests for 1 minute to illustrate the cold start time but also get a visual sense of request rate and variance in response time:

Again we have some quite striking results. AWS has the lowest cold start time of around 1.5 seconds, Google is next at 2.5 seconds and Azure again the worst performer at 9 seconds. All three systems then settle into a fairly consistent response time but it’s striking in these graphs how AWS Lambda’s significantly better performance translates into nearly 3x as many requests as Google and 10x more requests than Azure over the minute.

It’s worth noting that the cold start time for the stock functions is almost exactly the same as for my main test case – the startup is function related and not connected to storage IO.


AWS Lambda is the clear leader for HTTP triggered functions – on all the runtimes I’ve tried it has the lowest response times and, at least within the volumes tested, the best ability to deal with scale and the most consistent performance. Google Cloud Functions are not far behind and it will be interesting to see if they can close the gap with optimisation work over the coming year – if they can get their flat our response times reduced they will probably pull level with AWS. The results are similar enough in their characteristics that my suspicion is Google and AWS have similar underlying approaches.

Unfortunately, like with the .NET scenarios, Azure is poor at handling HTTP triggered functions with very similar patterns on show. The Azure issues are not framework based but due to how they are hosting functions and handling scale. Hopefully over the next few months we’ll see some improvements that make Azure a more viable host for HTTP serverless / API approaches when latency matters.

By all means use the above as a rough guide but ultimately whatever platform you choose I’d encourage you to build out the smallest representative vertical slice of functionality you can and test it.

Thanks for reading – hopefully this data is useful.

Azure Functions vs AWS Lambda – Scaling Face Off

Since I published this piece Microsoft have made significant improvements to HTTP scaling on Azure Functions and the below is out of date. Please see this post for a revised comparison.

If you’ve been following my blog recently you’ll know I’ve been spending a lot of time with the Azure Functions – Microsoft’s implementation of a serverless platform. The idea behind serverless appeals to me massively and seems like the natural next evolution of compute on the cloud with scaling and pricing being, so the premise goes, fully dynamic and consumption based.

The use of App Service Plans (more later) as a host mechanism for Azure Functions gave me some concern about how “serverless” Azure Functions might actually be and so to verify suitability for my use cases I’ve been running a range of different tests around response time and latency that culminated in the “real” application I described in my last blog post and some of the performance tests I ran along the way. I quickly learned that the hosting implementation is not particularly dynamic and so wanted to run comparable tests on AWS Lambda.

To do this I’ve ported the serverless blog over to AWS Lambda, S3 and DynamoDB (the, rather scruffy, code is in a branch on GitHub – I will tidy this up but the aim was to get the tests running) and then I’ve run a number of user volume scenarios against a single test case: loading the homepage. The operations involved in this are:

  1. A GET request to a serverless HTTP endpoint that:
    1. Loads 3 resources from storage (Blob Storage on Azure, S3 on AWS) in an asynchronous batch.
    2. Combines them together using a Handlebars template
    3. Returns the response as a string of type text/html.

On Azure I’m using .NET 4.6 on the v1 runtime while on AWS I’m using the same code running under .NET Core 1.0. It’s worth noting that latency on blob access remained minimal throughout all these tests (6ms on average across all loads) and when removing blob access from the tests it made little difference to the patterns.

Although the .NET 4.6 and Core runtimes are different (and accepted may exhibit different behaviours) these are the current general availability options for implementing serverless on the two platforms using .NET and both vendors claim full support for them. In Microsoft’s case some of the languages supported on the v1 Azure Functions runtime, the one tested here (v2 is in preview and has serious performance issues with .NET Core), are experimental and documented as having scale problems but C# (which runs under full framework .NET) is not one of them. Both vendors have .NET Core 2.0 support on the way and in preview but given the issues I’m waiting until they go on general availability until I compare them.

The results are, frankly, pretty damning when it comes to Azure Functions ability to scale dynamically and so let’s get into the data and then look at why.

A quick note on the graphs: I’ve pulled these from VSTS, it’s quite hard (or at least I don’t know how to!) equalise the scales and so please do look at the numbers carefully – the difference is quite startling.

Add 2 Users per Second

In this test scenario I’ve started with a single user and then added 2 users per second over a 5 minutes run time up to a maximum of 500 users:

We can see from this test that AWS matches the growth in user load almost exactly, it has no issue dealing with the growing demand and page requests time hover around the 100ms mark. Contrast this with Azure which always lags a little behind the demand, is spikier, and has a much higher response time hovering around the 700ms mark.

This is backed up by the average stats from the run:

It’s interesting to note just how many more requests AWS dealt with as a result of it’s better performance: 215271 as opposed to Azure’s 84419. Well over twice as many.

Constant Load of 400 Concurrent Users

This test hits the application with 400 concurrent users from a standing start and runs over a 10 minute period simulating a sudden spike or influx of traffic and looking at how quickly each serverless environment is able to deal with the load. Neither environment was completely cold as I’d been refreshing the view in the browser but neither had had any significant traffic for some time. The contrast is significant to say the least:

Let’s cover AWS first as it’s so simple: it quickly absorbs the load and hits a steady response time of around 80ms again in under a minute.

Azure, on the other hand, is more complex. Average response time doesn’t fall under a second until the test has been running for 7 minutes and it’s only around then that the system is able to get near the throughput AWS put out in a minute. Pretty disappointing and backed up by the overall stats for the run:

Again it’s striking just how improved the AWS stats over the Azure figures.

Constant Load of 1000 Concurrent Users

Same scenario as the last test but this time 1000 users. Lets get into the data:

Again we can see a similar pattern with Azure slow to scale up to meet the demand while with AWS it is business as usual in under a minute. Interestingly at this level of concurrency AWS also error’d heavily during the early scaling:

It should be noted that AWS specifically instructs you to implement retry and backoff handlers on the client which in the load test I am not doing, additionally at this point I am seeing throttle events in the logging for the AWS function – this is something I will look to come back to in the future. However its interesting to note the contrasting approaches of the two systems: Azure inflates it’s response time while AWS prefers to throw errors.

The average stats for the run:

Azure Functions

I don’t think there’s much point dancing around the issue: the above numbers are disappointing. Azure is slow to scale it’s HTTP triggered functions and once we get beyond the 100 concurrent users point the response times are never great and the experience is generally uneven. For customer facing API / web serving where low latency and response time are critical to a smooth user experience this really rules it out as an option. And it’s not just the .NET 4.6 variant that is poor as can be seen from my previous posts where I stripped test cases down to the most basic scenarios and used a variety of frameworks. The best case for Azure scaling I’ve found is using a CSX approach to return a string but even that lags behind AWS doing real work as the test cases in this post do:

using System.Net;

public static async Task<HttpResponseMessage> Run(HttpRequestMessage req, TraceWriter log)
    log.Info("C# HTTP trigger function processed a request.");

    var response = req.CreateResponse();
    response.StatusCode = HttpStatusCode.OK;
    response.Content = new StringContent("<html><head><title>Blog</title></head><body>Hello world</body></html>", System.Text.Encoding.UTF8, "text/html");

    return response;

With 1000 concurrent users over 5 minutes:

And with the add 2 users per second scenario:

Even in this final case, and remember this Azure Function is only returning a string, we can see the response time creeping up as the user load increases and the total number of requests served is only 77514 to AWS’s 215271 over the same period with a much lower number of requests per second.

In an additional attempt to validate my conclusion that the Azure Function system is poor at scaling I pointed the AWS Lambda installation at Azure Blob Storage instead of S3. In this test other than the function entry point semantics the code running on AWS is now taking exactly the same branches as the Azure tests and using the same underlying storage mechanism, albeit with a hop across the Internet to access the storage. I ran this scenario using the 400 concurrent user scenario:

We can see from this that other than a slightly increased response time due to the storage being hosted in another data centre AWS continues to perform well and scales up almost immediately and response time remains steady and low. We can also see their is no issue with Azure Blob Storage – if there was an issue there we’d expect to see it impact these results.

With these additional validation tests (an empty workload and AWS running against Blob Storage) that pretty much isolates the issue to the Azure Function runtime.

And it’s a shame as the developer experience is great, there is solid documentation, and plenty of samples, and the development team on Twitter are ludicrously responsive – to the point that I feel bad saying what I need to say here. I will reach out to them for feedback.

Why is this the case? Well I’d suggest the root of the issue is how the system has been built on top of App Service Plans. It’s not all that, well, serverless and you still find yourself worrying about, well, servers.

On Azure an App Service Plan is essentially a collection of rented servers / reserved compute power of a given spec (CPU, memory) and capabilities. Microsoft have layered what they call a Consumption Plan over this for Azure Functions which provides for automatic scaling and consumption based pricing. Unfortunately if you track what is going on your Functions are running on a limited number of these servers which you can evidence by tracking the instance ID and by sharing state between your functions (to be clear: this is not good!).

Essentially the level of granularity for scaling your functions remains, as in a traditional hosting model, at the server level and as your system scales up instances are slowly being added – but this is throttled tightly presumably to prevent Microsoft’s costs from spiralling out of control.

Now because they run on Application Service Plans you can switch hosting away from the Consumption plan onto a standard plan (which allows additional Azure features to be used) but this, to me, completely defeats the point of serverless. I’m paying for reserved compute again and managing server instance counts. I may as well not have bothered in the first place!

It’s hard to escape the feeling that Microsoft had to play catch up with AWS Lambda (it launched as a preview in late 2014 and went into general release in April 2015 whereas Azure Functions launched as a preview in March 2016 ) and built something they could market as serverless computing as quickly as they could by reusing existing compute and scaling systems on Azure.

Would I still use Azure Functions? Yes sure – in back end scenarios where latency isn’t all that important they’re a great fit. Anything that impacts user experience? No. Definitely not at this point.

It will be interesting to see if Microsoft revise the hosting model, I suspect if they do it’s some time off as currently they seem focused on the v2 runtime which isn’t a hosting change (as far as I can see) but rather giving Functions the ability to support more languages and .NET Core.

AWS Lambda

I’ll preface this by saying I am absolutely not an AWS expert so it’s harder for me to speculate about the underlying architecture of Lambda however… the numbers don’t lie: AWS manages to respond to changes in demand very quickly and, until I started to hit throttle limits (which I would need to speak to AWS Support to have lifted), is very consistent in response times.

I’ve not tried any state sharing but I would expect it to fail: it looks like Amazon have containerised at the Function level, rather than the host server, and this is what allows them to operate as you’d expect a serverless environment to. Both scaling and billing can then be at the function level.

Would I use AWS Lambda? Yes. But as most of my development work is on Azure I’m really hoping Microsoft bridge the capability gap.

Wrap Up and Next Steps

If you’ve followed this far – thanks! I’m a big fan of the serverless model but the Azure implementation of serverless looks like something of a compromised offering at this point and I’d be cautious of recommending it without understanding in detail the usage requirements as you will quickly hit choppy water.

I am planning on repeating similar experiments with the queue processing I began some time ago and if I get any information from Microsoft around this topic will make any corrections as appropriate. This is one of those times I’d love to have got things wrong.

Serverless Blog – Christmas 2017 Project

Happy New Year everyone – I hope everyone had a great break and has a fantastic 2018.

Much like last year I’d set some time aside over the Christmas break to tinker with something fairly left-field and somewhat experimental (algorithmic art) but unfortunately spent a lot of the break ill. This left me with a lot less time on my hands than I’d planned for and based my project around – I’d hoped to spend 4 to 5 days on it and an additional day for writing this blog post but had been left with only around 12 hours available for the implementation.

That being the case I scrabbled around for something smaller but still interesting and useful to me and that I thought would fit into the reduced amount of time I had available. I decided I’d attempt to put together a Minimum Viable Product for replacing my WordPress based blog with something that looks and feels the same to the reader but is entirely serverless in it’s architecture. My aim was to get, in no particular order, something that:

  • Renders using a similar look and feel to my current blog
  • Supports the same URL patterns for posts so that I could port my content, do a DNS change, and wouldn’t cause Google or linking sites a problem
  • Has super-cheap running costs
  • Has high uptime
  • Uses Markdown as it’s post authoring format
  • Has fast response times (< 100ms for the main payload)
  • Is capable of scaling up to high volumes of concurrent users
  • Support https for all content as my current blog does
  • Was deployed and running on an endpoint at the end of my allotted time (you can try it out here)

Knowing I only had 12 or so hours to spend on this I didn’t expect to be flicking the switch at the end of the second day and migrating my blog to this serverless system but I did want to have it running on my domain name, fairly sound, and be able to prove the points above with a working Minimum Viable Product. From a code quality point of view I wanted it to be testable and reasonably structured but I wasn’t aiming for perfection and expected low to zero automated test coverage.

The challenge here was covering enough ground in 12 hours to demonstrate an MVP worked and was in a sufficiently developed state that it was clear how the quality could be raised to a high degree with a fairly small amount of additional work.

If you’re interested in seeing the code it can be found on GitHub. If you use this as a basis for your own projects please bear in mind this was put together very rapidly in just over 12 hours – it needs more work (see next steps at the bottom of this post).


Normally when I undertake a project like this I’ve had the chance to roll it around in my head for a few days and can hit the ground running. With the late change of direction I didn’t really get the chance to do that and so I really came into this pretty cold.

As I wanted to replace my current WordPress blog with a serverless approach a good place to start seemed to be looking at it’s design and my workflows around it. The layout of my blog is pretty simple and every page has the same structure: a title bar, a content panel, and a sidebar:

In addition their are only really 4 types of page: a homepage made up of the most recent posts, posts, category pages, and archives. The category and archive pages simply list the posts within a category and month respectively. The only thing that causes site change is the addition or editing of a post that can cause all those pages to require update.

I do most of my writing on the train and use the markdown format which I subsequently import into WordPress for publishing. This means I don’t really use the editing capabilities of WordPress (other than to deal with markdown to WordPress conversion issues!) and so was comfortable simply uploading the Markdown to a blob container for this serverless blog. This left the question of how to get any metadata into posts (for example categories) and I decided on a simple convention based approach where an optional block of JSON could be included at the start of a post. That way that too could be maintained in a text editor.

Given all that my general approach (at this point best catgorised as a harebrained scheme) was to render the components of the site as static HTML snippets using a blob triggered Azure Function and assemble them into the overarching layout when a user visits a given page with page requests being handled by HTTP triggered Azure Functions – one per page type. I toyed with the idea of going full static and re-rendering the whole website on each update but felt this “mostly” static approach revolving around the side components might provide a bit more flexibility without much performance impact as all I’m really doing to compose a page is stitching together some strings, and if I were to actually start using this I’d like to add a couple of dynamic components.

In any case having settled on that approach I mapped the architecture out onto Azure services as shown below:

In addition to using Azure Functions as my compute platform for building out the components I picked a toolset I’m either working with day to day or have used in the past:

  • C# and .NET Core
  • Visual Studio 2017 and Visual Studio Code
  • Handlebars for page templating
  • Blob and Table storage

I briefly considered using CosmosDB as a datastore but my query needs were limited and it would bump up the running cost and add complexity for no real gain and so quickly discounted it.


With the rough planning complete it was time to knuckle down with the laptop, a quiet room, a large quantity of coffee, and get started on some implementation. Bliss!

In order to make this readable I’ve organised my approach into a linear series of steps but like most development work there was some to-ing and fro-ing and things were iterated on and fleshed out as I moved through the process.

My general approach on a project like this is to prioritise the building out of a vertical slice and so here that meant starting with a markdown file, generating enough of the static assets that I could compose web pages, and then a couple of entry points so I could try it out in the Azure environment.

Step 1 – Replicating the Styling of my Existing Blog

As this project is really about markdown in and HTML out I wanted to start by ensuring I had a clearly defined view of that final output and so I began by creating a HTML file and CSS file that mimicked the layout of my existing blog. Design is always easier for a none-designer when you have a reference and so I quite literally opened up my current blog in one tab, my candidate HTML file in another tab and iterated over the content of it and the CSS until I had something that was a reasonable approximation.

While I’m not going to pretend that the CSS is a stunning piece of artistry this didn’t take long and I was sufficiently in the ballpark after just an hour.

Time taken: 1 hour

Step 2 – Creating a Solution and Code Skeleton

Next up was creating a solution skeleton in Visual Studio establishing the basic coding practices along with the models I expected to use throughout. My previous work with Azure Functions has been for small and quite isolatable parts of a wider system rather than being the main compute resource for the system and so I’d not really had to think too hard about how to organise the code.

Something I knew I wanted to carry over as a pattern from my previous work was the concept of “thin” functions. The function methods themselves are, to me, much like actions on a ASP.Net Core / Web API controller – entry points that accept input and return output and should be kept small and focused, handing off to more appropriate implementers that are not aware of the technicalities of the specific host technology (via services, commands etc.). Not doing that is a mix of concerns and tightly ties your implementation to the Functions runtime.

While I wanted to separate my concerns out I also didn’t want this simple solution to inflate into an overly complex system and so I settled on a fairly traditional layered approach comprised, from an implementation point of view, of 4 assemblies communicating over public C# interfaces but with fully private implementations all written to .NET Standard 2.0:

  • Models – a small set of classes to communicate basic information up and down, but not out of (by which I mean they are not persisted in a data store directly nor are they returned to the end user), the stack
  • Data Access – simple implementations on top of table storage and blob storage
  • Runtime – the handful of classes that do the actual work
  • Functions – the entry point assembly

Mapped out this ultimately gave me a solution structure like this:

The remaining decision I needed to make was how to handle dependency injection. An equivalent system written with, say, ASP.Net Core would use an IoC container and register the configuration during startup but that’s state that persists for the lifetime of the server and functions are ideally stateless. Spinning up and configuring an IoC container for each execution of a function seemed needlessly expensive so I made the decision to use a “poor mans” approach to dependency injection with the Runtime and Data Access assemblies each exposing a static factory class that was responsible for essentially implementing the “Resolve” method for each of my instantiable types and that exposed public create methods for the public interfaces of each layer.

For the limited number of classes I have this approach worked pretty well and allowed me to write testable code in the same way as if I was using a fully fledged container.

Time taken: 1 hour

Step 3 – Creating the Layout, Posts and the Homepage

The first step in turning my earlier HTML and CSS work into something that could be used to create a real blog from real posts was to write a Handlebars template for the overall layout that could stitch together the main content and sidebar into a full HTML document. Based on my earlier work this was pretty simple and looked like this:

        <link href="{{stylesheetUrl}}" rel="stylesheet" />
        <link rel='stylesheet' href='' type='text/css' media='all' />
        <link href="{{faviconUrl}}" rel="shortcut icon" type="image/x-icon" />
        <div class="title-panel">
            <div class="container">
                <a class="primary-title" href="/">{{blogName}}</a>
        <div class="container">
            <div class="content">            
                <div class="reading">                            
                <div class="sidebar">
        <div class="footer-panel">
            <div class="container">
                Copyright &copy; {{defaultAuthor}}

Along with this I created a pair of methods in my composition class to bring the components of the site together:

public async Task<string> GetHomepage()
    return await GetWrappedContent(() => _outputRepository.GetHomepageContent());

private async Task<string> GetWrappedContent(Func<Task<string>> contentFunc)
    Task<string> templateTask = _templateRepository.GetLayoutTemplate();
    Task<string> sidebarTask = _outputRepository.GetSidebar();
    Task<string> contentTask = contentFunc();

    await Task.WhenAll(templateTask, sidebarTask, contentTask);

    string template = templateTask.Result;
    string content = contentTask.Result;
    string sidebar = sidebarTask.Result;

    TemplatePayload payload = new TemplatePayload
        BlogName = _blogName,
        DefaultAuthor = _defaultAuthor,
        PageTitle = _blogName,
        ReadingContent = content,
        Sidebar = sidebar,
        StylesheetUrl = _stylesheetUrl,
        FavIconUrl = _favIconUrl
    Func<object, string> compiledTemplate = Handlebars.Compile(template);

    string html = compiledTemplate(payload);
    return html;

To generate posts I needed to read a post from an IO stream and then convert the Markdown into un-styled HTML and for that I used the excellent CommonMark.NET which I hid behind an injected helper to facilitate later testing. After conversion the post is saved to the output blob store:

Post post = await _postRepository.Get(postStream);
string html = _markdownToHtmlConverter.FromMarkdown(post.Markdown, post.UrlName, post.Author, post.PostedAtUtc);
await _outputRepository.SavePost(post.UrlName, html);

Actually deserializing the post took a little more effort as I needed to also parse out the metadata and this can be seen in the PostParser.cs implementation.

The homepage on my blog is basically the most recent n posts compiled together and so to do this I used another Handlebars template:

{{#each this}}
    {{#if @index}}
        <div class="post-spacer"></div>

To order the posts on the homepage (and later the sidebar) I need to track the “posted at” dates of each post. I can’t use on the LastModified property of the blob as that won’t deal with updates correctly and to migrate my content over I need to be able to set the dates as part of that process. To do this I persisted some basic data to an Azure Storage table.

And finally I created a handlebars template for generating a hard coded sidebar based on my sample.

Time taken: 3 hour

Step 4 – Blob Triggered Post Processing Function

At this point I had a bunch of code written for processing markdown and generating web pages but no way to call it and so the next step was to implement a function that would listen for new and updated blobs and generate the appropriate assets:

public static class ProcessPost
    public static async Task Run([BlobTrigger("posts/{name}", Connection = "BlogStorage")]Stream myBlob, string name, TraceWriter log)
        log.Info($"ProcessPost triggered\n Blob Name:{name} \n Size: {myBlob.Length} Bytes");


        IStaticAssetManager staticAssetManager = Factory.Instance.GetRenderer();
        await staticAssetManager.AddOrUpdatePost(myBlob);

This function demonstrates the use of some of the principles and practices I thought about during the first step of this process:

  • The Azure Function is small and restricts it’s actions to that domain: it takes an input, sets up the subsequent environment and hands off.
  • The poor mans dependency injection approach is used to resolve an instance if IStaticAssetManager.

I tested this first locally using the Azure Functions Core Tools and other than some minor fiddling around with the local tooling it just worked which I verified by checking the output blob repository and eyeballing the contents. No great genius on my part: I’m using things I’ve used before and am familiar with to solve a new problem.

Time taken: 1 hour

Step 5 – Homepage and Post Functions

Next up was to try and render my homepage and for this I wrote a new function following the same principles as before:

public static async Task<ContentResult> Run([HttpTrigger(AuthorizationLevel.Anonymous, "get", Route = "home")]HttpRequest req, TraceWriter log)
    log.Info("C# getContent HTTP trigger function processing a request.");

    IResponseRenderer responseRenderer = Manager.Factory.Instance.GetResponseRenderer();
    string content = await responseRenderer.GetHomepage();

    return new ContentResult
        Content = content,
        ContentType = "text/html",
        StatusCode = 200

This worked but I encountered my first challenge of the day: the function was on a path of which is not going to allow it to function as the root page for my website. In fact if I went to the root I would instead see the Azure Functions welcome page:

While this is a perfectly fine page it’s not really going to help my readers view my content. Fortunately Azure Functions also include a capability called Proxies which allow you to take any incoming request, reshape it, and call an alternate backend. I had no idea if this would work on a root path but wrote the simple pass through proxy shown below:

  "$schema": "",
  "proxies": {
    "HomePageProxy": {
      "matchCondition": {
        "route": "/",
        "methods": [
      "backendUri": "https://%BlogDomain%/home"

That matches on a GET request to the root and sends it on to my home page handler. This works absolutely fine when run on Azure but doesn’t work locally in the Core Tools – they seem to use the root path for something else. I need to do more investigation here but for now, given it works in the target environment and I only have 12 hours, I settled on this and moved on.

To remove the api component of the URI on my future functions I also modified the hosts.json file used by Azure Functions setting the HTTP routePrefix option to blank:

  "http": {
    "routePrefix": ""

Writing this I’m wandering if this is what’s causing my issues with the root proxy on the local tools. Hmm. Something to try later as I can accomplish the same with another proxy.

Time taken: 2 hours

Step 6 – Load Testing

With my homepage compositor function written and a working system deployed to the cloud with this first fully representative vertical slice I wanted to get a quick handle on how it would cope with a reasonable amount of load.

Visual Studio Team Services is great for quickly throwing lots of concurrent virtual users against a public endpoint. I set up a test with a fairly rapid step up in the number of users going from 0 to 400 concurrent users in around 5 minutes and then staying at that level for another 15 minutes.

I knew from my casual browser testing that the response from the homepage function for a single user page load on a quiet system took between 60 and 100ms which I was fairly pleased about. I expected some divergence from that as the system scaled up but for things essentially to work.

Much to my surprise and horror that was not the case. As the user count increased the response time started running at around 3 to 4 seconds per request and generated an awful lot of errors along the way. The system never scaled up to a point where the load could really be acceptably dealt with as can be seen below:

I blogged about this extensively in my last post and so won’t cover it again here but the short version is that the Azure Functions v2 .NET Core runtime (that is still in preview) was the culprit. To resolve things I migrated my functions over to .NET 4.6.2 and after doing so and running a similar test again I got a much more acceptable result:

Average response time over the run averages 700ms and the system scaled out pretty nicely to deal with the additional users (and I pushed this up to 600 on this test). The anecdotal experience (me using the browser with the cache disabled as the test ran) was also excellent and felt consistently snappy throughout with timings of between 90ms and 900ms with the majority that I saw taking around 300ms (it’s worth noting I’m geographically closer than the test agents to the Azure data centre the blog is running in – VSTS doesn’t run managed agents from UK South currently).

As part of moving to .NET 4.6 I had to make some changes to my functions, an example of this is below:

public static async Task<HttpResponseMessage> Run([HttpTrigger(AuthorizationLevel.Anonymous, "get", Route = "home")]HttpRequestMessage req, TraceWriter log)
    log.Info("GetHomepage triggered");

    IWebPageComposer webPageComposer = Factory.Instance.GetResponseRenderer();
    string content = await webPageComposer.GetHomepage();

    HttpResponseMessage response = req.CreateResponse(HttpStatusCode.OK);
    response.Content = new StringContent(content, Encoding.UTF8, "text/html");            

    return response;            

Time Taken: 3 hours

Step 7 – Sidebar Content

To maintain a sidebar I needed to maintain some additional metadata – what posts belong in what categories which I’m pulling from the (optional) JSON annotation of the Markdown files I outlined earlier. An example of that can be seen below:

    createdAtUtc: '2017-12-29 10:01:00',
    categories: [
    urlName: 'aUrlNameForAPost',
    author: 'James Randall'

The categories get parsed into a very simple table storage class:

internal class CategoryItem : TableEntity
    public string UrlName => PartitionKey;

    public string PostUrlName => RowKey;

    public string DisplayName { get; set; }

    public string PostTitle { get; set; }

    public DateTime PostedAtUtc { get; set; }

    public static string GetPartitionKey(string categoryUrlName)
        return categoryUrlName;

    public static string GetRowKey(string postUrlName)
        return postUrlName;

The UrlName‘s referenced above are just (by default) camelcase alphabetic strings used to identify posts as part of a URI and as such are unique (within the context of a blog). Because all this activity takes place on the backend and away from user requests I’ve not bothered with any more complex indexing strategies or further storage tables to store the unique set of categories – instead when I need to organise the categories into a hierarchical structure or get the category names I simply load them all from table store and run some simple LINQ:

internal class CategoryListBuilder : ICategoryListBuilder
    public IReadOnlyCollection<Category> FromCategoryItems(IEnumerable<CategoryItem> items)
        var result = items.GroupBy(x => x.UrlName, (k, g) => new Category
            UrlName = k,
            DisplayName = g.First().DisplayName,
            Posts = g.OrderByDescending(x => x.PostedAtUtc).Select(x => new PostSummary
                PostedAtUtc = x.PostedAtUtc,
                Title = x.PostTitle,
                UrlName = x.PostUrlName
        }).OrderBy(x => x.DisplayName).ToArray();

        return result;

This is something that might need revisiting at some point but this isn’t some uber-content management system, it’s designed to handle simple blogs like mine, and, hey, I only have 12 hours!

I take a similar approach to generating the list of months for the archives section of the sidebar and then create it as a static asset with a Handlebars template:

<h2>Recent Posts</h2>
    {{#each recentPosts}}
        <li><a href="/{{urlName}}">{{title}}</a></li>        
    {{#each archives}}
        <li><a href="/archive/{{year}}/{{month}}">{{displayName}}</a></li>
    {{#each categories}}
        <li><a href="/category/{{urlName}}">{{displayName}}</a></li>

Time taken: 2 hours

Step 8 – Wrap Up

With most of the system working and problems solved all that was left was to fill in a couple of the empty pages: post lists for categories and archives. The only new code I needed to this was something to summarise a post, for the moment I’ve taken a quick and dirty approach to this based on how my content is structured: I look for the title and the end of the first paragraph in the HTML output.

And having got that again I simply use another Handlebars template to generate the output and a couple more functions to return the content to a user.

Time taken: 2 hours

Conclusions and Next Steps

Did I succeed? Well I have my MVP,  it works, and it ticks off what I wanted! However I took 14 hours to put this together rather than the 12 I’d allowed. Most of the overrun was due to the performance issue with .NET Core and the Azure Functions v2 runtime, it took a little while to pin down the cause of the issue as the starting point for my investigation was based on the (generally reasonable!) assumption that I’d done something stupid.

Given that and as it’s New Year I’m going to give myself a pass and class this as a resounding success! A few takeaways for me:

  • Azure Functions are very flexible and serverless can be a great model but there are some definite limitations in the Azure Function implementation some of which stem from the underlying hosting model – I’m going to come back to this in a future blog post and, time allowing, contrast them with AWS Lamda’s.
  • Implementing this using Azure Functions was not really any harder than using ASP.Net Core or Web API.
  • Never underestimate the need to test with some load against your code.
  • It’s always spending some time on identifying the main challenges in a project and focusing your efforts against them. In this case it was covering enough ground quickly enough to validate the design without making things a nightmare to move on and into a more professional codebase.
  • If you really focus its amazing how much you can get done quickly with modern tools and technologies.
  • Development is fun! I had a great time building this small project.
  • Blogging takes even longer than development. The real overrun on this project was the blog post – I think its taken me the best part of 2 days.

If I continue with this project the next steps, in a rough priority order, will be to:

  1. Add unit tests
  2. Introduce fault tolerance strategies and logging
  3. Add a proper deployment script so others can get up and running with it
  4. Test it with more content (extracted from my blog)
  5. Improve code syntax highlighting
  6. Ensure images work

Finally the code that goes along with this blog can be found over on GitHub:

Azure Functions v2 Preview Performance Issues (.NET Core / Standard)

I’ve been spending a little time building out a serverless web application as a small holiday project and as this is just a side project I’d taken the opportunity to try out the new .NET Core based v2 runtime for Azure Functions and the new tooling and support in Visual Studio 2017.

As soon as I had an end to end vertical slice I wanted to run some load tests to ensure it would scale up reliably – the short version is that it didn’t. The .NET Core v2 runtime is still in preview (and you are warned not to use this environment for production workloads due to potential breaking changes) so you would hope that this will get fixed by general release but right now there seem to be some serious shortcomings in the scalability and performance of this environment rendering it fairly unusable.

I used the VSTS load testing system to hit a single URL initially with a high volume of users for a few minutes. In isolation (i.e. if I run it from a browser with no activity) this function runs in less than 100ms and normally around the 70ms mark however as the number of users increases performance quickly takes a serious nosedive with requests taking seconds to return as can be seen below:

After things settled down a little (hitting a system like this from cold with a high concurrency is going to cause some chop while things scale out) average request time began to range in the 3 to 9 seconds and the anecdotal experience (me running it in a browser / PostMan while the test was going on) gave me highly variable performance. Some requests would take just a few hundred milliseconds while others would take over 20 seconds.

Worryingly no matter how long the test was run this never improved.

I began by looking at my code assuming I’d made a silly mistake but I couldn’t see anything and so boiled things down to a really simple test case, essentially the one that is created for you by the Visual Studio template:

public static IActionResult Run([HttpTrigger(AuthorizationLevel.Anonymous, "get", Route = null)]HttpRequest req, TraceWriter log)
    log.Info("C# HTTP trigger function processed a request.");

    var result = new OkObjectResult("hello world");

    return (IActionResult)result;

I expected this to scale and perform much better as it’s as simple as it gets: return a hard coded string. However to my surprise this exhibited very similar issues:

The response time, to return a string!, hovered around the 7 second mark and the system never really scaled sufficiently to deal with a small percentage of failures due to the volume.

Having run a fair few tests and racking up a lot of billable virtual user minutes on my credit card I tweaked the test slightly at this point moving to a 5 minute test length with step up concurrent user growth. Running this on the same simple test gave me, again, poor results with average response times of between 1.5 and 2 seconds for 100 concurrent users and a function that is as close to doing nothing as it gets (the response time is hidden by the page time in the performance chart below, it tracks almost exactly). The step up of users to a fairly low volume eliminates the errors, as you’d expect.

What these graphs don’t show are variance around this average response time which still ranged from a few hundred milliseconds up to around 15 seconds.

At this point I was beginning to suspect the Functions 2.0 preview runtime might be the issue and so created myself a standard Functions 1.0 runtime and deployed this simple function as a CSX script:

using System.Net;

public static async Task<HttpResponseMessage> Run(HttpRequestMessage req, TraceWriter log)
    var response = req.CreateResponse();
    response.StatusCode = HttpStatusCode.OK;
    response.Content = new StringContent("hello world", System.Text.Encoding.UTF8, "text/plain");

    return response;

Running the same ramp up test as above shows that this function behaves much more as you’d expect with average response times in the 300ms to 400ms range when running at 100 concurrent users:

Intrigued I did run a short 5 minute 400 concurrent user test with no ramp up and again the csx based function behaved much more in line with what I think are reasonable expectations with it taking a short time to scale up to deal with the sudden demand but doing so without generating errors and eventually settling down to a response time similar to the test above:

Finally I deployed a .NET 4.6 based function into a new 1.0 runtime Function app. I made a slight mistake when setting up this test and ramped it up to 200 users rather than 100 but it scales much more as you’d expect and holds a fairly steady response time of around 150ms. Interestingly this gives longer response times than .NET Core for single requests run in isolation around 170ms for .NET 4.6 vs. 70ms for .NET Core.

At this point I felt fairly confident that the issue I was seeing in my application was due to the v2 Function runtime and so made a quick change to target .NET 4.6 instead and spun up a new v1 runtime and ran my initial 400 concurrent user test again:

As the system scales up, giving no errors, this test eventually settles at around the 500ms average request per second mark which is something I can move ahead with. I’d like to get it closer to 150ms and it will be interesting to see what I can tweak so I can on the consumption plan as I think I’m starting to bump up against some of the other limits with Functions (ironically resolving that involves taking advantage of what is actually going on with the Functions runtime implementation and accepting that its a somewhat flawed serverless implementation as it stands today).

As a more general conclusion the only real takeaway I have from the above (beyond the general point that it’s always worth doing some basic load testing even on what you assume to be simple code) is that the Azure Function 2.0 runtime has some way to go before it comes out of Preview. What’s running in Azure currently is suitable only for the most trivial of workloads – I wouldn’t feel able to run this even in a beta system today.

Something else I’d like to see from Azure Functions is a more aggressive approach to scaling up/out, for spiky workloads where low latency is important there is a significant drag factor at the moment. While you can run on an App Service Plan and handle the scaling yourself this kind of flies in the face of the core value proposition of serverless computing – I’m back to renting servers. A reserved throughput or Premium Consumption offering might make more sense.

I do plan on running these tests again once the runtime moves out of preview – I’m confident the issue will be fixed, after all to be usable as a service it basically has to be.

Azure Functions – Expect Significant Clock Skew

While running the experiment I posted about on Sunday I annotated the message I sent on to the event hub with the Azure Functions view of the current time (basically I set a property to DateTime.UtcNow) and, out of curiosity, grouped my results by second long tumbling windows based on that date. This gave me results that were observably different than when doing the same with the enqueue date and time logged by the Event Hub (as an aside there is some interesting information about Event Hubs and clock skew here). My experiments didn’t need massively accurate time tracking as I was really just looking for trends over a long, relative to the clock skew, period of time however I looked at some of the underlying numbers and became suspicious that there was a significant degree of clock skew across my functions.

I reached out to the @AzureFunctions team on Twitter asking how they handled clock sync on the platform and one of the engineers, Chris Anderson, replied confirming what I suspected: there are no guarantees about clock sync on the Azure Functions platform and, further, you should expect the time skew to be large.

That means you can’t really obtain a consistent view of “now” from within an Azure Function. You could go and get it from an external source but that in and of itself is going to introduce other inaccuracies. Essentially you can’t handle dynamic time reliably inside a function with any precision and you’re limited to working with reference points obtained upstream and passed in.

Definitely something to be aware of when designing systems that make use of Azure Functions as it would be easy to use them in scenarios applying timestamps expecting some sense of temporal sequence.

This isn’t, of course, a new problem – dealing with precise and accurate time in a distributed system is always a challenge and requires careful consideration but it does underline the importance of understanding your cloud vendors various runtime environments.

Azure Functions – Queue Trigger Scaling (Part 1)

I’m a big fan of the serverless compute model (known as Azure Functions on the Azure platform), but in some ways its greatest strength is also its weakness: the serverless model essentially asks you to run small units of code inside a fully managed environment on a pay for what you need basis that in theory will scale infinitely in response to demand. With the increased granularity this is the next evolution in cloud elasticity with no more need to buy and reserve CPUs which sit partially idle until the next scaling point is reached. However as a result you lose control over the levers you might be used to pulling in a more traditional cloud compute environment – it’s very much a black box. Using typical queue processing patterns as an example this includes the number of “threads” or actors looking at a queue and the length of the back off timings.

Most of the systems I’ve transitioned onto Azure Functions to date have been more focused on cost than scale and have had no particular latency requirements and so I’ve just been happy to reduce my costs without a particularly close examination. However I’m starting to look at moving spikier higher volume queue systems onto Azure Functions and so I’ve been looking to understand the opaque aspects more fully through running a series of experiments.

Before continuing it’s worth noting that Microsoft continue to evolve the runtime host for Azure Functions and so the results are only really valid at the time they are run – run them again in 6 months and you’re likely to see, hopefully subtle and improved, changes in behaviour.

Most of my higher volume requirements are light in terms of compute power but heavy on volume and so I’ve created a simple function that pulls a message from a Service Bus queue and writes it, along with a timestamp, onto an event hub:

[return: EventHub("results", Connection = "EhConnectionString")]
public static string Run([ServiceBusTrigger("testqueue", Connection = "SbConnectionString")]string myQueueItem, TraceWriter log)
    log.Info($"C# ServiceBus queue trigger function processed message: {myQueueItem}");
    EventHubOutput message = JsonConvert.DeserializeObject<EventHubOutput>(myQueueItem);
    message.ProcessedAtUtc = DateTime.UtcNow;
    string json = JsonConvert.SerializeObject(message);
    return json;

Once the items are on the Event Hub I’m using a Streaming Analytics job to count the number of dequeues per second with a tumbling window and output them to table storage:

    System.TimeStamp AS TbPartitionKey,
    '' as TbRowKey,
    SUBSTRING(CAST(System.TimeStamp as nvarchar(max)), 12, 8) as Time,
    COUNT(1) AS totalProcessed
TIMESTAMP BY EventEnqueuedUtcTime

For this initial experiment I’m simply going to pre-load the Service Bus queue with 1,000,000 messages and analyse the dequeue rate. Taking all the above gives us a workflow that looks like this:

Executing all this gave some interesting results as can be seen from the graph below:

From a cold start it took just under 13 minutes to dequeue all 1,000,000 messages with a fairly linear, if spiky, approach to scaling up the dequeue rate from a low at the beginning of 23 dequeues per second to a peak of over 3000 increasing at a very rough rate of 3.2 messages per second. It seems entirely likely that this will go on until we start to hit IO limits around the Service Bus. We’d need to do more experiments to be certain but it looks like the runtime is allocating more queue consumers while all existing consumers continue to find their are items on the queue to process.

In the next part we’re going to run a few more experiments to help us understand the scaling rate better and how it is impacted by quiet periods.


  • If you're looking for help with C#, .NET, Azure, Architecture, or would simply value an independent opinion then please get in touch here or over on Twitter.

Recent Posts

Recent Tweets

Invalid or expired token.

Recent Comments




GiottoPress by Enrique Chavez