Automatically translate this page?
1 May, 2019

April 2019 change log

Major flespi improvements in April 2019

In terms of weather, this April was the best from the past ten Aprils that I remember. But despite the warm and sunny environment, the flespi team was able to concentrate on the product and deliver a wonderful set of new features. 

But before I list them, let me tell a few words about our monthly uptime which was as low as 99.925%. This is the lowest we had since 99.70% in September 2017 and particularly contrasting to the super-stable March with 100% uptime. In April our uplink provider performed major works on our network - initially flespi and Wialon networks were located in the same VLAN segment and as our product grows, we decided to split both networks into separated VLAN segments to gain full independence from Wialon datacenter. This operation was performed on Sunday, April 7 and according to our NOC and our health checking bots the world was twice disconnected from flespi — for 844 and 1043 seconds. At the same time, from the software point of view, we didn’t have any problems with the backend, so if we exclude these maintenance works, theoretically our monthly uptime could be 100% again. But what doesn’t kill us makes us stronger — this was the last maintenance we need this year to grow further. And it’s over. So I believe next months will be much better and we will try to reach as high as 99.995% yearly uptime or close to it.

  • One more thing about flespi locations around the world — we are coming to Russia. As you may know, Russia is heading towards regional Intranet by restricting access to resources outside of the country and issuing special laws. However, flespi has a few customers in Russia and Gurtam, in general, dominates the Russian market with more than 40% of commercial vehicles equipped with GPS connected to its flagship product — Wialon. That’s why we decided to build the first secondary flespi location in Moscow. Once the new data center goes in production within a few months, it will be possible for our users to select the location of their flespi items — devices, channels, streams, even subaccounts, and MQTT sessions. No matter where the current location of the item is, they will work in the same cloud, so it is possible to have one channel in the Netherlands, another channel in Moscow, and devices in Moscow as well — such setup will operate seamlessly through both channels. As for the further expansion of our data centers network, we are currently considering the North American region and are already testing the data center in Canada for stability and customer service.
  • Our Analytics engine as a tool to do reports and events notification system received thorough documentation in KB and a dedicated all-in-one blog article. In May we intend to install the automated analytics system that will be generating and storing intervals in real-time. The engine is already quite stable and functional even for manual (not-automated) operations (e.g. to run reports over devices messages). So we invite everybody to try it and give feedback. One of our most active flespi users based in Brazil has already developed the system using flespi Analytics, substituted its past implementation, and got 100x performance boost at a lower cost because in cloud systems you are paying for CPU time and in flespi Analytics you are paying only for storage.
  • One of the largest GPS tracking systems in the world with 150K+ devices is being migrated to flespi at the moment. The new version of the platform will incorporate such flespi technologies as Analytics, HASD and will heavily rely on the flespi MQTT Broker and especially its MQTT 5.0 protocol implementation that gives extra possibilities for developers. Until now, they have already activated 35K devices on flespi running in parallel to the original platform and you can monitor project stages at our platform status page. Specifically for this user, we adapted flespi.io panel GUI to work seamlessly with 50K devices. And it works really smooth thanks to the power of HASD and MQTT.js. Have you ever imagined such quantities in the web application before?
  • We finally released MQTTTiles application capable of visualizing data coming from MQTT on shareable and flexible dashboards. We are adding more widgets and features but you may already use it in your projects. The shared dashboard is mobile-friendly and the application is quite lightweight, so it should perfectly run even on older smartphones. It means that instead of developing own dashboard-like interfaces, you may just publish the data calculated by your backend application into MQTT (as our automated analytics engine will soon do) and deliver stable and good-looking apps with minimal time overhead to your customers. This is our alternative to Sensolator (for those who work with Wialon and heard about it).
  • Regarding MQTT itself, one of our developers — Alexander Kiranov — has released open-source MQTT 5.0 compatible client written in pure Lua. So currently the following three open-source MQTT 5.0 client libraries are being maintained by Gurtam developers:
  • If you are working with flespi, I suggest you try any of these using our MQTT API and feel the power of an alternative approach to software design.
  • In April we integrated three new protocolstopflytech, coban, and xexun and trying to understand all these cheap Chinese devices suddenly discovered the following relation: all cobans are xexuns, but not all xexuns are cobans +)
  • Since we are diving deeper into protocols integrations in our daily work we decided to automate some operations and now proactively monitor what is happening in our channels and reacting to parsing errors even before our users notice them. It means that if you are connecting the device to flespi and for some reason, it does not work today, please retry tomorrow — chances are that our clever monitoring bot detected the issue, handled the problematic packets, informed us about the case, and we might have already fixed it.
  • In May Gurtam participates in a range of exhibitions. You are welcome to talk with our representatives about flespi as well. Do not miss a chance to meet us on May 7-9 in Mexico, on May 14-16 in Johannesburg, on May 21-23 in Sao Paolo
  • And if you are in GPS Tracking or Fleet Management business, don’t miss the opportunity and join the largest Telematix conference in Minsk on July 30 - August 1 and meet everybody from the flespi team in person there.

Flespi is still a team with 99% of engineers. That’s why in May, despite all these events, we will continue working on the product itself. We’ll be launching the new location in Moscow and will install the automated Analytics engine — the one that will obviously change the vision of how simple and efficient telemetry data processing can be. The one that will be able to answer the following question in seconds: which of these millions of vehicles visited this geofence in the last five years and when.



Get the latest updates and monthly newsletters from flespi in your inbox

1 July, 2019 | change log | Aliaksei Shchurko

June 2019 change log

Major flespi improvements in June 2019

3 June, 2019 | change log | Aliaksei Shchurko

May 2019 change log

Major flespi improvements in May 2019