Azure Stream Analytics Reviews

We are using Azure Stream Analytics for small to medium size streaming datasets where you would like to flag patterns from the stream. It works well or pairs well with IoT edge scenario use cases that are on Azure. If you have exceptional conditions, such as a sensor being way off the average for the last one to five hours, then you can flag a scenario. It works well with the IoT infrastructure that Azure provides.

We didn't end up using Azure Stream Analytics in production, or for a client, we implemented it. However, Azure Stream Analytics is something that you can use to test out streaming scenarios very quickly in the general sense and it is useful for IoT scenarios. If I was to do a project with IoT and I needed a streaming solution, Azure Stream Analytics would be a top choice.

The most valuable features of Azure Stream Analytics are the ease of provisioning and the interface is not terribly complex.

Azure Stream Analytics could improve by having clearer metrics as to the scale, more metrics around the data set size that is flowing through it, and performance tuning recommendations.

I have been using Azure Stream Analytics for approximately three months.

Azure Stream Analytics is stable.

Azure Stream Analytics can improve the scaling and the connectivity to external datasets.

We are not using this solution extensively and we do not plan to increase usage.

The level of support quality depends on how much you purchased.

I rate the support from Azure Stream Analytics a four out of five.

Positive

The initial setup of Azure Stream Analytics was straightforward. It has a quick startup time and is easy to start.

I did the implementation of Azure Stream Analytics for my client. We have the developer setting the solution up and once it's in production, your infrastructure team can monitor it just like any other solution. Since it's Azure, it has a lot of metrics that allow you to be proactive to flag an issue if there is one.

I have seen a return on investment with Azure Stream Analytics. If you're not doing terribly complex scenarios, this is a quick and fast way to have your streaming pipeline set up. You won't have to invest a lot into its deployment because it's the cloud. You are not paying any upfront capital.

I rate the price of Azure Stream Analytics a four out of five.

I have evaluated other solutions, such as Databricks

Azure Stream Analytics it's good for proofs of concept and for scenarios that are not too complex. It's promising in the future, but if you start to scale out, you might want to consider other scaling solutions, such as Databricks.

Got it. And do you see a return on investment with this one?

I rate Azure Stream Analytics a nine out of ten.

We were doing some level of stream data processing, so we had some use cases which were related to IoT. We had some IoT devices getting data in from other IoT devices and Azure Streaming Analytics has a special streaming analytics offering for IoT devices. Basically it was used for that. 

I basically use two features that are useful. One is Azure Event Hubs, and that is used in conjunction with Azure Streaming Analytics. One is the broker and one is the processing engine. With the processing engine, the SQL way of dealing with things, with streams, is what I like, compared to other solutions, which are more like Scala or Spark-based, where you need to know the language. This was comparatively easy to use with its ability to write SQL on streams.

Technical support is pretty helpful. 

It's my understanding that the setup is pretty straightforward.

With Azure specifically, the drawback is it is a very Azure-specific product. You can't connect it to external things out of Azure. For example, Spark or Databricks can be used in any cloud and can be used in AWS. This product doesn't work that way and is very Azure-specific. It's not a hybrid solution and it's not a cloud-agnostic solution, where you put it on other clouds, et cetera. 

We had some connections which we wanted to make with AWS, which we couldn't do with this. We had to use something else for that.

Early in the process, we had some issues with stability.

You cannot do joins on streams of data. For example, one stream joining with another stream. Real-time to real-time joins, you're not able to do that. You can only join your stream with static data from your Azure storage. 

I've used the solution for one and a half to two years.

There were some issues with the IoT jobs when streaming Azure Streaming Analytics, which are high proof now. That said, earlier, we used to have a lot of issues with the erratic behavior of jobs. If data is not in the way they expect it, if they are not modeled correctly, then the jobs tend to break or fail quite a lot. That was one issue we had.

We've been in touch with technical support. There was a time when jobs failed a lot and we couldn't understand the reason. When we talked to the spec tech support, they've looked into our data and told us that it's not exactly modeled as how Azure Stream Analytics needs it. That wasn't very clear when we got it. 

They were helpful. There were issues which they handled, which they told us about. The communication was great.

We had the support package included.

I'm now an analyst, so I don't use the products per se, however, prior to this, I have used Azure Streaming Analytics quite a lot. Currently, I'm working a bit on Databricks Spark Streaming. These two are, I would say, what I have used personally.

The product was set up before I started out, however, what I can say, having set up some things personally, is it is comparatively straightforward and the Microsoft support on that is comparatively good.

In terms of pricing, you can't compare it to open source solutions. It would be higher compared to open source, of course, however, with the support and everything you're getting, I would say the price, in general, is fair. 

I have seen AWS as well and can compare it to that and I would say it is fair. The problem is it is not exactly dynamic or serverless, with how the way things are in the cloud. Therefore, it is not completely utilized. You have to set up things beforehand with some level of capability and capacity beforehand. In regards to the price, it's not too high and also not too low.

Their pricing is not exactly serverless. It's per hour. A lot of others are moving towards pricing based on the amount of data you pull. Streaming Analytics charges per hour, and in that sense, you need to set up the capacity by trial and error, literally. 

I'm comparing the Azure Stream Analytics, AWS Kinesis, GCP Pub/Sub, and Dataflow. So I'm currently in the process of writing that research.

If you are in the Azure world completely, and you're using the Microsoft stack completely, and you do not have the need to go in any other cloud, then it makes sense to use this solution as it integrates very well within the Azure ecosystem. 

For IoT use cases, if you want to do real-time dashboarding with Power BI, it's great. Those kinds of things are where it has its niche. However, if you want a cloud-agnostic kind of solution, where you do not want to be stuck with just Microsoft, then there are other solutions out there such as Confluent, Kafka, Spark Streaming with Databricks, et cetera. You'll get the flexibility you need using any of those platforms.

I'd rate the solution at a seven out of ten. We had some issues with the jobs not behaving properly. They promise a lot, however, sometimes that doesn't happen and we realized that later. Some things under the hood, we couldn't really understand and we needed to get in touch with support. Those kinds of issues are where I would say it needs a bit of improvement, and maybe that's why I cut off two or three points.

We use the solution to analyze application logs. It helps review incidents and production issues.

The solution's most valuable feature is its ability to create a query using SQ.

The solution's interface could be simpler to understand for non-technical people. Also, the chart feature could be more user-friendly. Presenting or elaborating on an incident to management or executives from other non-technical departments becomes challenging. We have to create another graph and include it in the presentation slides.

We have been using the solution for two years.

The solution is stable.

The solution is scalable. We have around more than 100 developers using it.

We follow virtual documentation in case of technical errors for the solution.

If you have Azure DevOps or are using the Azure ecosystem, you must go for the solution. I rate it an eight out of ten.

We mainly use this product for real-time data injection and reporting purposes.

We find the query editor feature of this solution extremely valuable for our business.

Also, the multiple window types that are available in this solution, are extremely helpful.

The solution doesn't handle large data packets very efficiently, which could be improved upon.

We would also like more variation in the output types that this solution can produce.

We have been using this solution for three or four years.

The stability of this solution was problematic initially, but it has improved as later versions have been released, and is now very good.

We have found this product to be easily scalable.

The initial setup of this solution is very easy, and the deployment takes place instantly.

The licensing for this product is payable on a 'pay as you go' basis.  This means that the cost is only based on data volume, and the frequency that the solution is used.

This solution is easy to use and has good fault tolerance, so I would recommend it to other organizations.

I would rate this solution an eight out of ten.

The company I'm working for is basically one of the biggest companies in the entire Gulf region, including Dubai, Qatar, and Oman. Our core domain is providing logistics. They have different warehouses across the country, and we use it to track the movement of forklifts and people working at the warehouses. 

The main thing we are focusing on right now is accident avoidance. For example, one forklift is coming through one aisle, and another person is trying to enter the same aisle. We provide a solution that can track the person and forklift in real-time. 

We're also using it to solve many business problems one by one. Stream Analytics plays a major role in streaming all these huge datasets because we have warehouses spread across the country. It's able to handle millions and millions of events in a few seconds.

If we're not using Stream Analytics, how can we track the real-time? From the end user's perspective, Stream Analytics forms the main backbone for the entire pipeline and all the technologies.

The company can see the real-time location and track everything, just because of Stream Analytics. Without Azure Stream Analytics, we can't do any real-time tracking. We can use other messaging systems like SQL, but when it comes to scaling, collecting, getting a lot of events, recalling it, find out where it's used, Stream Analytics is better. You might have to collect from millions and millions of services and devices and beacons. All of that would be pushing the data into the Stream Analytics. 

I like the IoT part. We have mostly used Azure Stream Analytics services for it. This is the most valuable part because this is using the streaming service. It's valuable because there's no other way for us to handle it. 

It has the support of Azure storage, long storage, and access data. It has support from the SQL server. Azure also supports added access to data because we need the data from static data and dynamic data to represent them, and that's not going to change very frequently.

When we are getting the warehouse's location through this stream analytics, we have to merge some information from our static database, and finally, we have to show it all within a dashboard or something on the map.

Without the streaming analytics part, I don't think it's possible to handle it. We can use some other messaging system, but we might have some scaling issues and among others too. I know that Stream Analytics is fantastic in that we don't have to worry about any other activities. We can further scale it too. We can go for the upgraded service if needed, based on our traffic and the number of data we have been receiving.

Natively, I found it beneficial, and the integration was smooth. We're already using some other Microsoft technology packs, so it's easy to integrate them all.

As the Stream input and output enables very smooth integration with other cloud services, for example, Azure Cloud Concepts, or Cosmos DB, with minimum coding, and with the minimum level of queries, we can directly output all these outputs and push the normal data for historical data storage.

The collection and analysis of historical data could be better. We use historical data and an assimilating algorithm to give us insights into the entire business process. 

We can collect all the historical data periodically to get insights into current business trends. For example, which area is getting emptied most of the time or which area is getting underutilized, and so on. 

I've been working with Azure Stream Analytics for about two years.

We don't have to worry about scalability. It's in the cloud and can have millions and millions of things connected. The software part is easy to scale. You just have to add all the hardware. For the web application, the hosting part can be scaled. We don't have to worry about the desktop as the solution is deployed in the cloud. The scalability is based on our choices. It's not like it's manually hosted in private, and we have to scale it vertically.

Our infrastructure team has the flexibility to call the Microsoft guys to look into the matter if there is something wrong on their part.

The initial setup was very complex because of the hardware. We had to spend almost an entire day just to put the hardware part in the right places, following some best practices. 

It took us more than one and a half years, and we're still left with some deployments to do. We initially tested it in a few small areas, and then we expanded it to cover the entire area.

I found it a little challenging, we struggled, and we did it. We're still doing a lot of stuff for the elite features and other deployments. We follow the deployment strategy, and it's almost automated. We're trying to add a few features and deploying them. The final stage of deployment is where the rest of the entire process is done through continuous integration.

It requires maintenance in terms of hardware and the software part. I don't think any solution is totally bug-free. We generate service requests all the time, and they are fixing it.

The IoT hardware requires more maintenance because we know we have limited battery life, and we have to check all the devices. We need to keep checking those things, and we have automated that. But it still needs to be manually reconnected to the battery.

We have a team of people supporting this project. We have about ten members, some of whom were core developers. Four or five developers developed the cloud part. Two hardware engineers were responsible for all these deployments in the warehouse.

I would advise potential users to properly plan and structure their static data and the reference data before putting it into the Stream Analytics.  

On a scale from one to ten, I would give Azure Stream Analytics an eight.

I used it once for a project demo to a customer for an IoT solution. In this demo, the data was collected from the sensors, and it was sent to Power BI reports. The collected data was analyzed by using the analytics tools to get some insights.

This project was the first project for our company to start the development of IoT solutions. We have only used it for a demo, and we have kept it for demo for other customers. If any customer wants to deploy it, we would use it in production.

Real-time analytics is the most valuable feature of this solution. I can send the collected data to Power BI in real time.

It is not complex, but it requires some development skills. When the data is sent from Azure Stream Analytics to Power BI, I don't have the access to modify the data. I can't customize or edit the data or do some queries. All queries need to be done in the Azure Stream Analytics. 

I have used this solution for a few months.

I don't know about its stability because we didn't use it in production. We only used it for testing.

Its scalability is okay. In Azure Stream Analytics, I can add more data sources through reference or IoT hub. 

For the demo, we had a team of 20 users. The customer was looking at allowing around 20,000 users for this solution.

I contacted their technical support once because I found an issue with Azure Stream Analytics. The technical support engineer was very supportive.

The initial setup was straightforward for me. I read some articles on the Internet, and it worked fine for me. It took us one to two weeks to deploy it.

If you want to deploy IoT services, this solution will be very helpful for real-time applications and for collecting data.

I would rate Azure Stream Analytics a seven out of ten. 

We have different kinds of IoT devices placed in different countries including the UK, US, and others. They are configured with our IoT hub and we get the logs from them accordingly.  We have these logs connected with the Stream Analytics suites and Microsoft Power BI. Whatever updates and other activity is happening on the devices are streamed into Azure and Power BI so that we can see them.

If we find any error messages then we have to check the health of the corresponding IoT devices, databases, and configuration.

This gives us a real-time monitoring system that we can use to analyze the health of our IoT devices. Previously, when something was not working properly then we would receive messages in our email using the TeamWork application. Now, instead of checking email, we receive an alert ping that we can hear, which allows us to evaluate how well the machine is doing. We can check the performance and other relevant metrics.

In general, it gives us more visibility in terms of what is going on. We used to receive between 10,000 and 20,000 emails per week, which was hectic for us to calculate and keep track of. Since implementing Azure, we have been able to monitor things very easily. Not only does it create an interval for the logs but it reduces the number of duplicates.

We have not eliminated the messages that come in as email, as high-priority messages are still delivered in that manner. For example, if there is a power shut-down then we will be notified via email. This is set up in case we miss these types of messages in the BI platform.

The most valuable features are the IoT hub and the Blob storage. All of the logs and other data that we are getting can be stored in Blobs.

The interface is user-friendly.

There may be some issues when connecting with Microsoft Power BI because we are providing the input and output commands, and there's a chance of it being delayed while connecting.

I have been using Azure System Analytics for just more than one year.

This product is stable but if our VM goes down then we are not able to get a proper instance update. When this happens, we need to kill these instances. Situations like this only happen rarely.

The scalability is based on the requirements. If the requirements are high then highly-scalable machines are needed. If it is more manageable then it is cheaper. I think that scaling is really about the cost.

We have a development team and an operations team that is working with Azure Steam Analytics. There are seven or eight people in the operations team. The customer also has access to the platform if they require it.

If you raise a ticket with technical support then they will contact you within 24 hours. However, we have not faced many issues, so we haven't had much involvement with them.

There is a diagnostic tool available in Azure and you can check to see if you have any issues on your end. If there are problems then you can contact support for assistance.

Overall, I think that the support is very helpful.

Since transitioning from our email-only solution, we have been able to set the interval that we use to retrieve logs from the devices.

We did not use a similar product before this one for the same purpose. The company has been using Azure since before I joined, although they had used AWS for other tasks. At this company, I have not had the opportunity to work on AWS.

I have not completed a deployment for production purposes. Rather, I have performed a setup for training with Azure and an IoT simulator. In this case, we just check the logs during my practice session. My role in the operation was to lead the management team.

The training deployment that I completed was user-friendly and anyone can easily do it. Even as part of the operations team, I was able to capture the details and complete the deployment really quickly.

The only difficulty that I faced was connecting with the different machines in the outside layer, such as BI or Kibana. Depending on the application I was connecting with, there were issues with it.

The deployment was done by our development team, and they are responsible for the maintenance as well. Because it is a platform as a service, Azure takes care of almost everything.

I am not familiar with the details of the investment. This is something that is handled completely by the product owner. This would be my manager or the Delivery Manager.

The cost of this solution is less than competitors such as Amazon or Google Cloud. If we only use one hour then we are only charged for one hour. It is very easy and some products are more expensive.

Azure Stream Analytics is something that we were able to easily learn. It doesn't take much programming sill, so I feel that it is easy to start using.

Other than the problem with delays in connecting to Microsoft BI, Kibana, or other monitoring tools, I don't have any other issues with this product.

I would rate this solution a nine out of ten.

Our primary use case is mainly to ingest real time data streams into permanent storage places like databases, block storage, etc.

The biggest improvement for us has been that it now takes much less time for us to receive valuable information. Basically, as soon as it appears in our real time data source, in a matter of seconds, it is already in our database.

The value of this solution is the deep integration it provides with other Azure resources which we use a lot. Our whole infrastructure is pretty much based on Azure so ease of integration is a valuable feature. Secondly, the simplicity of the solution is great. You don't need to set up much, you just make a selection, select a destination, and you're off. 

There are some improvements that could be made, first of all in pricing, because right now the pricing is a bit unclear. It's hard to estimate how much of that is a local issue but you can't figure out how prices are calculated or the proprietary part of the cost. Another area that could be improved is that if something does go wrong, it's very hard to investigate what caused it and why. The logging is available but it lacks detail and doesn't provide much information.

I've been using this solution for two years. 

The solution has an acceptable level of stability although, as mentioned, if it does fail, it's pretty difficult to find out the cause. 

It's very easy to scale this solution. We probably have a couple of hundred users and we have developers who deal with maintenance. This is our main tool for real time data streaming. 

The initial setup is quite straightforward. Because of the good integration, you select your real time data, store the destination where you want to write it and you probably don't even need to transform with data. You basically create a mapping descent source. We had a proof of concept in place, so I would say deployment took two working days without having a deployment plan. 

We have a good ROI because we are able to deliver solutions very quickly and customers are happy with that. 

We evaluated and carried out a comparison with Oracle. The results were pretty much the same for both in terms of real time data streams, but were very much tied to their own cloud solutions. If you work with Oracle i'ts probably best to go with Amazon.

My simple advice would be to not scale up initially. Also, if you have questions don't just rely on the official documentation, but use other resources such as a blog by a developer, because sometimes that can be more helpful than documentation provided by the company.

The best advice I can offer would be that if there is a simple solution available, do not try to complicate things. 

I would rate this solution an eight out of 10.