This time I don’t have much to say about the weather although it is currently in its most beautiful season. I will leave the description of picturesque October for next year or till the moment when we have less information about product development. But this time October is different from other months mostly due to the following factors:
We had the worst-ever platform uptime in the last 32 months;
We introduced a great number of new powerful features to the platform;
Let’s start with uptime which was 99.949% in October. There was a coincidence of two factors: scheduled maintenance and a DDoS attack with unscheduled maintenance to mitigate it. The whole story which started on the evening of October 10 you can read in our NOC. As a result during that week, we were down for 1,377 seconds in total, from which 1,014 seconds are due to a DDoS attack and 363 seconds are due to scheduled maintenance and required changes we required in the network in order to mitigate the attack. The attack by itself was not directly to the flespi network, but towards another platform hosted nearby. For us, it was the final motivation to change the network subsystem and, as result, we migrated our whole network into dedicated IPv4 address space. This is what I planned a few years earlier and actually, we were fully prepared for this migration, but it was unclear if we needed it and in what form. As it turns out — yes, we needed it. And now we are much more diverse and protected.
That change also triggered a few more changes as indicated in the Telematics Hub changelog:
Our default gateway IP form which the flespi network accesses outside world services has changed to 185.213.2.10. This may require some additional whitelisting on your side.
Our primary IP for channels has changed to 185.213.2.30. At the same time, we continue providing services on the legacy IP 193.193.165.37. It’s in our monitoring system, we receive automatic notifications about its accessibility and we maintain its operation as usual. What has been changed is the uptime tracking status — now we do not count the unavailability of this IP address as platform downtime as we do with the default address 185.213.2.30 used by channels. In case of any serious issue with this address we will indicate it into NOC, but still not count that time as downtime.
By the way, the worst month ever in our history was September 2017 with 99.70% uptime. We just started flespi and had a lot of changes in the infrastructure which resulted in such a low uptime. You may read this five-year-old story here, just to compare things in flespi now and back then.
Well, that was the bad part of the news. The good part is we did a lot of new things on the platform.
- The very first new feature I want to share with you is device groups. Groups are not only about simple devices grouping and visualizing. Groups are our automation tool for you in order to simplify the creation and maintenance of complex flespi configurations. With groups, you can perform the smart assignment of devices between streams, calculators, and plugins and enhance access control via tokens.
- Next is embedded access to billing information directly in flespi.io. Now our commercial users can control payment methods and automatic CC charging configuration, update the list of emails for billing notifications, and download invoices. And since in flespi everything is designed by developers and for developers, you can access and manage the billing data via REST API as well if needed.
- We love engineers and complex solutions. But even complex solutions are better managed with simple tools. Here is a nice and powerful sandbox for plugin development you can enjoy. It's an open-source tool, so feel free to use it directly or fork. This is a great thing for those who believe in the power of data transformation and need something special from the flespi.
- Another super simple, but extremely powerful feature is now available in msg-expression plugin expressions and in the “validate_message” field of all plugins — the possibility to reference previous (known) parameter values. For example, you want to detect the moment when the driver's door opened or the vehicle has just stopped or started. You can now use “din.1 && (din.1 != #din.1)” type of expression for that. The possibility to address the exact change between two consecutive messages on the plugin level opens a whole new range of applications. The only thing you should be aware of is if your device sends messages in backward order, this functionality will be useless as flespi can only operate with the values of parameters it currently knows.
- Two new stream types are now available for our users — one to integrate data from flespi into the Mapon platform and another to Redlist. I counted seven public and known GPS tracking platforms you can already use flespi for data ingestion and transformation. It means that any piece of equipment integrated with flespi can automatically be forwarded to these seven well-known platforms. Sounds like an efficient opportunity for hardware manufacturers to spread their devices. And for our clients, it's an opportunity to use multiple platforms seamlessly and in parallel with whatever functionality they need in their specific projects.
- For those who missed it, I recommend reading about this corporate rent-a-car project implemented on top of flespi which got first place in the “public transport” category of IoT Project of the Year. As far as I know, this is a reusable project so if you have a flespi account and need a nice Uber-like system for your clients you are welcome to contact KloudIP.
And a list of ongoing changes you may expect in flespi soon:
First of all, new pricing is in effect from November 1st. Just another reminder for those who missed it a month ago.
Only one month left until we cease support of channel commands functionality. Here is the article with the information on the upcoming change. In order to make migration smoother we increased the maximum available pending commands in the device commands queue from 20 to 100.
And we are changing the telemetry publication mode (MQTT). Starting from November 23 only real latest parameters will be published. In general, this is what users expect from us and if your devices are sending data in the “oldest first” mode, you will not be affected.
Quite a lot of new exciting features, isn’t it? We will not be able to maintain such a pace of new releases from month to month and I think in the next few months we are about to stabilize and prepare for the next jump. Our next large feature is the NoCode functionality in flespi. This is going to be a module that eliminates the need for our users to maintain specific-purpose lambda functions outside flespi. We are still thinking about how to implement it in the simplest way but at the same time make it flexible enough for versatile applications. Once released, if everything is implemented correctly, our users will be able to save thousands of USD per month on the cloud infrastructure and even more on engineering staff. With a few mouse clicks, you will be able to implement functions that previously required weeks of development.