1 February, 2024

January 2024 change log

Major flespi improvements in January 2024.

The first month of 2024 marked a significant success for flespi. Within just one month, our users registered over 50,000 new devices on the platform. Around 35,000 devices were registered due to some major migrations or new customer acquisitions by our commercial users, while the remaining 15,000 reflects a natural, organic growth spread widely among our 4,000+ users.

For us, this monthly growth is a new milestone, representing a more than 6% increase in less than a month. Now, only 140,000 devices stand between us and our goal of reaching the first million. The team is thrilled with this achievement and extends a big thanks to our users for growing with us. As always, we'll do whatever we can to deliver the best telematics backend as a service in the world.

Our January uptime was 99.9941%. Due to a software bug, our REST API services dealing with devices and channels experienced CPU overload for about 3 minutes. Not all nodes, but noticeable parts of them, with 50% of REST API requests towards devices and channels, could fail due to some timeout during this problem. We identified and fixed the issue almost immediately.

We released the msg-cancel plugin, capable of preventing device message registration. The plugin can be used to skip unnecessary messages from being injected into the message registration pipeline.

We added the option to enable and disable tokens and subaccounts. This feature targets our users managing projects with complex architectures, with requirements to block/unblock certain users, for example, due to non-payment.

For our video telematics direction, we finally provisioned the infrastructure to implement video codecs conversion on the fly. Now, both Howen and Streamax can store and transmit video data in the highly efficient and compact H.265 (HEVC) codec, while flespi will re-encode video data to the highly adopted and compatible H.264 codec.

We also fully support both live streaming and video file download functionality for the new Jimi IoT (Concox) video telematics line - JC181 and JC450. For Queclink CV100LG and CV200 devices, we implemented video download functionality.

From our initial plan for video telematics features, only the support for RTMP live video stream remains, which we will integrate soon. This integration will enable live streaming on Queclink and legacy Jimi IoT dashcams. We also have intentions to integrate at least one new video device manufacturer with a high-end type of hardware to broaden the choice of supported manufacturers suitable for any type of project.

And last but not least is our new comprehensive video telematics integration guide. We invite hundreds of local Telematics Service Providers of all sizes with their platforms already using flespi to do this integration and offer your users a full set of video telematics features. It has never been so easy to compete with large platforms, especially in such specific segments. The integration itself can be completed within just a few days, and your profit on this type of installation per a single device/vehicle is much higher than what you can get from a standard vehicle tracking system.

To increase the transparency of flespi, we extended the storage period for logs from 30 to 365 days for all entities (channels, devices, streams, MQTT, REST API access logs, etc.).

We implemented raw traffic storage for webhooks and plugins again to increase transparency and simplify any kind of troubleshooting.

Let me remind you that we do not charge our users for traffic and logs storage. Actually, they are widely used and priceless when we need to diagnose any issue that you may have with devices or streams. Now webhooks and plugins are also in this club.

We released an overview of top manufacturers for us in 2023. Not too many changes since 2022 for leading positions, but in the middle of the rank, the competition is fierce.

We constantly refresh information in our Knowledge Base, from which our AI Assistant draws information for its answers. One interesting article that I want to highlight is this one that explains how to detect and inform you immediately about some event, like a trip that has just finished. This article explains a lot about how to mix calculators and webhooks functionality for such kinds of solutions.

And finally, we published the overview and explained the architecture of codi - our new AI assistant. Codi was warmly welcomed by our users, and one by one, they started using it in their daily tasks - mostly to consult on how to efficiently apply flespi functionality to solve their problems. It is really effective in this type of consulting, as it knows everything we’ve posted to our KB and Blog.

The effect of AI in our communication is amazing. It handled 30% of the communication between us and our users in 24/7 mode and saved us and them a huge amount of time due to low latency and high quality of its replies.

We decided to continue this track and train it to configure the most popular devices, diagnose all kinds of problems you may have, check raw traffic in case of device error, and create a comprehensive report for the manufacturer in case the problem is in the device or towards us in case of a protocol implementation problem, generate REST API calls and PVM code for you upon request, and so on. This is a long track, but I firmly believe that generative AI services are not just hype. We saw it already with a 30% decrease in support efforts for an engineer - a great way to go.

Later in February, I will post our roadmap for 2024 for you to sync your upcoming plans with ours. Stay tuned!