2 September, 2019

August 2019 change log

Major flespi improvements in August 2019

August in Belarus is the peak month for vacations. And despite the fact that six out of nine members of the flespi team had a 1+ week vacation this August, we produced a lot of interesting stuff.

  • The month itself started with the TelematiX conference held by Gurtam in Minsk. This year's conference had more than 500 attendees from all over the world and featured quite a few products and projects based on flespi. The progress is clearly visible — last year a lot of people told they would be doing flespi-based developments and now we see that some of them already have them implemented. For the flespi team, the best motivation is to see how our technologies help and serve people’s needs. During the conference, Gurtam also organized a special hackathon for developers led by Gurtam Labs lead developer — Aleksey Shmigelski and one of the best proofs of efficiency of using flespi in telematics applications is the fact that three out of four ready-to-use applications created at Hackathon are based on flespi.
  • Platform uptime for August was very close to four nines — 99.9891%. We experienced a few downtimes due to scheduled maintenance — we migrated our PostgreSQL meta-data databases for all entities identifiers from 32-bit to 64-bit. All our team members are very excited about flespi monthly uptime numbers each time trying to keep it as close to 100% as possible. And once we had the first 245 seconds of outage for meta-data manipulation operations over telematics hub entities (channels, devices, streams creation/deletion, etc) due to database patching, the guys decided to run all the patches by the end of August. Having analyzed the consequences of the first patch, we merged other metadata patches much better and caused just a 2-second downtime. We decided to move even further and migrated PostgreSQL cluster from 9.6 version to latest 11 with two small interruptions of 45 seconds in total. At the same time, the monthly uptime for our telematics data storage and MQTT Broker services was still 100%.
  • We finally launched own forum and invite our users to share their thoughts about flespi, ask questions and of course help newcomers. We are also using the forum as a comprehensive changelog for any kind of activity we are doing with the platform, including all modifications to protocols and integration of new devices. It’s fairly easy to subscribe to any tag or specific topic to keep yourself informed about any changes. It’s even more important if you rely on flespi in any kind of production project.
  • In August we reached our target for the beginning of 2018 — we reached the mark of 30+ Commercial plan customers. The initial goal was set at the end of 2016 and the actual subscriptions latency vs. expected was just 8 months. Not too bad as we didn’t have any kind of statistics based on which we could do more precise planning. Now we target 50 Commercial plan users by the beginning of 2020. Among our users, there are a lot of startups and a few well-known brands delivering telematics platforms and operating with hundreds of thousands of devices. This exactly covers the market segments we initially planned and I’m glad that reality correlates with our expectations.
  • Important reminder — this week (September 4) we will remove deprecated REST API methods for manipulating devices configuration. Also, we are going to deprecate abque storage element soon. We preliminary publish all such changes in our NOC, this is just a friendly reminder if you are not yet subscribed there.
  • We have integrated the skypatrol protocol with the enhanced support for both Skypatrol and Enfora devices. Also, we integrated streaming from flespi telematics hub in special so-called nis protocol. This format is widely used for inter-platforms communications in Russia.
  • We introduced the changes to the mqtt protocol configuration in order to support any kind of payload (not only valid JSON) as a message payload. This should enhance the usage of flespi MQTT channels to fetch the data from arbitrary systems and inject them into normal flespi telematics message pipeline.
  • A small tip that can save you a lot of time — now it is possible to open the context menu for any entity in flespi.io panel by long tap or right mouse click. The one handy item there that you may need is called Duplicate — to create new entities with similar configuration.
  • Our frontend team delivered a few interesting open-source components that can be used in web applications. One of them is FormBox that can take arbitrary JSON schema and render GUI for data input corresponding to this schema. We use it a lot to standardize the creation of any kind of user input. Another one is MapView that can be used to render the map with various markers, polylines or polygons. A nice sample of using MapView in our MQTT Tiles dashboard on top of flespi analytics is published on our forum — the visualization of flights in light aviation for the current flight season.

The flespi team spends most of the time on internal things — we are modifying the scheme of our services to operate them in multiple data centers simultaneously. The results of this work are not visible outside and we plan to be occupied with this activity until the end of the year.

But as usual, we will integrate more features and devices to currently supported telematics protocols and will help our users achieve their business goals with our technologies. Now you can track our activity on the forum — no need to wait for the changelog at the end of the month.

You can discuss this blog post on the forum as well.