2 November, 2017

October 2017 change log

Major flespi improvements in October 2017

October was a killer month! We made a few giant leaps toward our strategic goals and added a lot of features for you.

  • Our automated downtime detection system calculated 697 seconds of total platform downtime in October. This transparently converts to the high mark of 99.974% uptime. It was mostly our scheduled updates that caused the downtime — we had three major and a few minor upgrades, one of which required reverting to the previous version of the gateway system. Thanks to our automatic and robotized configuration system this process is easy but still takes some time to detect the problem and make a rollback decision. To minimize service interruption during upgrades, we reduced the services check timeout from 10 seconds to 1 second, and now upgrades should go smoother.
  • We received the first post on Gurtam forum in the flespi section— it relates to this 20-second downtime period during the upgrade. With our recent fixes, such problem should shrink to a few seconds. Anyway, we liked this case, and it's cool that there are people like us who require quality and constant uptime in this world where time-to-market for most products is becoming more important than their quality.
  • We participated in the GITEX exhibition in Dubai and discovered that excitement about flespi is growing exponentially among hardware manufacturers and telematics service providers, even those using modern feature-rich fleet management software products.
  • We transparently converted all our prices from EUR to USD, published the public offer on the website and received the first payment from a customer. At the time of October change log publication we are numbering 370 registered users, 200 of which are trying and using flespi platform inside own solutions.
  • We integrated three protocols: Systems and Technology, Baltic Car Equipment, ATrack and updated Wialon IPS protocol to transparently support its second version (Wialon IPS v2).
  • We created a nice FAQ for developers with use-cases and information on how to take advantage of flespi REST API.
  • In our storage system, we introduced containers parameters field — it has the same purpose as Telemetry feature in registry devices. With this feature and fields parameter, it is possible to request only a specific set of telemetry parameters from flespi containers. We decided to connect flespi containers and devices as a data source to open-source dashboard tool grafana that can create nice charts of telemetry change over time. Internally we are using our database system to store the platform operational statistics over time and use grafana to visualize it.
  • We implemented the initial integration of flespi gateway into Traccar in close cooperation with the Traccar founder. In November we plan to finalize the integration on a more complex level and are looking into other popular fleet management systems to charge them with power and flexibility of the flespi platform.
  • One of the most crucial development works this fall was the first version of our MQTT broker. We are going to move events processing inside flespi platform architecture into the MQTT-based bus to make it easier for internal consumers to subscribe to this or that event. We installed MQTT broker on multiple servers, gave it a name — mbus — and switched flespi streams processing from durable HTTP REST requests to MQTT. So far everything is perfect, and our implementation of MQTT broker is already faster than any open-source analog we tested. We plan to offer flespi platform users MQTT interface with PUB/SUB QoS up to 1 in December.
  • During integration flespi into other systems, we found out that current implementation of streams that works on a per-device level is not enough. We saw demand for push-like service for multiple devices, channel or even multiple channels. That's why we decided to redesign streams and make them a dedicated item in the registry. We will eliminate Wialon IPS and abque streams but will introduce stream based on Wialon Retranslator, so that multiple devices can send messages inside one TCP connection. So the initial list of protocols for updated streams will include: HTTP REST, MQTT (thingsboard.io and AWS formats) and TCP (Wialon Retranslator). We will upgrade existing streams where we can and delete streams we couldn't. These changes are coming in November.
  • November is also the time we are planning to finish the very first step in one of our strategic endeavors — the creation of a universal configurator for devices — generic, with the same interface for all types of protocols, powerful, implementing all settings specified by manufacturers in their protocols specifications. It took us six months and a few internal iterations to design and implement this feature, and I strongly believe that in a year or two this will be a number-one tool in the hands of most successful telematics system providers worldwide. So please subscribe to our flespi news Telegram channel or Facebook to be the first to try this tool in the field.
  • And last, but not least. Although software development process in a clean and warm office is a good thing, we never forget our origins — where our predecessors worked a few hundred years ago. So in October we implemented a real solution on a real farm and mounted temperature sensors on the cows. This dirty, sometimes bloody, but real-world work was a cool experience, especially for the IT-guys. Please read more in a dedicated article.

These were all the significant advancements we completed in October. We have two more months to craft some great Christmas and New Year gifts for you. So let’s keep in touch.