2 October, 2017

September 2017 change log

Major flespi improvements in September 2017
Disclaimer: flespi registry module merged with the gateway module to form the telematics hub between devices and platforms.

Dear flespi users!

Autumn is the harvest time, and we’ve reaped quite a few fresh goodies for you in September.

  • One of the most exciting activities that flespi participated in September was obviously Mobile World Congress exhibition in San-Francisco. The results of this event showed us that we are currently tapping into a new and unknown market segment. Being a pioneer is always a hard endeavor, but at the same time, it is a perfect opportunity to change the world for the better. And we are doing our best to achieve this.
  • One of the most remarkable technical achievements that September brought was the release of the second module for the flespi platform — the registry that provides high-level API for operating with GPS trackers and IoT devices as well as the streaming possibility to integrate them into various third-party platforms. As a part of this release, we also reorganized https://flespi.com website pages, so that they reflect the updated product structure.
  • We are going into production on some platforms that will rely on flespi API. We started to sign contracts, but right before signing, we did our best to check how the platform performs under load. We simulated all possible situations on all the components in our architecture — e.g., parts of PostgreSQL database cluster going offline, as well as power outages and freezing of our primary network gateway. Well, it was a very informative week that showed us a lot. Again, the results proved the concept that if you want it good, do each piece of software by yourself. Parts of architecture that were developed by us — the telematics database, even automatic administration system — they all behaved well. But standard open source components, such as the pacemaker, postgre, iptables and even Debian Linux kernel core — they all were the weak links. Anyway, we discovered many potentially shaky spots and fixed most of them. We also upgraded the operating system on all our servers from Debian 8 to Debian 9. The downside of all this testing and simulations was our uptime number. Our automatic calculation system computed only 99.70% uptime in September.
  • As a part of going into high-grade production, we created a separate Telegram channel — flespi NOC, where we post all technical incidents that may occur and information about updates. Plus our special flespi bot automatically posts there the downtime detection info — beginning of downtime with possible reason, its end, and the total duration after platform can respond and pass all tests.
  • We prohibited any unspecified fields in JSON data of each REST API request. According to the specification of JSON schema, we made “additionalProperties: false” by default, which prevents any unsolicited JSON parameters in API requests.
  • We introduced the latest telemetry property for devices. Latest telemetry represents the latest known values of device’s state. Each new message with the latest timestamp received by the device will update (with a tiny delay) the telemetry field of the device. So now it is possible to select devices that are offline for a certain time or stay in particular geographical rectangle with a specified trigger activated.
  • We also released an open source GUI of GPS tracking system based on the flespi registry. It is just the first step — we continue adding features to it, so you are welcome to join the community on Github and take part in its development.
  • We renamed some protocols and removed “-1” suffix, e.g., “teltonika-1” became “teltonika.” Also for each protocol that we support management commands, we now show them on the Commands tab on its page.
  • We added support for two new protocols — cguard (Russia) and ERM (Israel).
  • Our open source python adapter capable of reading messages from gateway got a few fixes for AWS MQTT retransmitter implementation. We also implemented an HTTP repeater onto custom HTTP server, so that messages from gateway can be used not only in pull mode via REST API but also via a push from this adapter.

As you can see, we are trying to get better both technically and regarding communication, proactively eliminate danger zones and become more user-friendly. We are always open to feedback from you, so please share your ideas or concerns with us.

Stay in the know of our latest developments at the flespi news Telegram channel.