The last improvements on our platform

The last improvements on our platform

Brice Vercoustre No comment Oct 13, 2015 11:00:00 AM Product information Lire en français

For a few months now, we have been steadily growing. The number of users and, as a consequence, the number of videos have increased the traffic on the videos hosted on our platform. 

Day by day, we work on the technical infrastructure to propose the best service possible to all our users. 

 Our servers improved

Any action realized by a user on Libcast (a video uploaded, a thumbnail changed, a publication...) sends a query on the Libcast Cloud infrastructure. Our users' queries were received on servers that are called "load balancers". Then, theses load balancers randomly split the queries on different frontend servers. 

Example : The user A wants to send a video, and the user B wants to change his video thumbnail. The two queries are received on a load balancer server. This one will send, randomly, the user A's query on the server #3 and the user B's query on the server #6. 

Some of our clients use the API to send a group of heavy videos on our platform, which could slightly slowdown the infrastructure and your browsing. 

Some queries, coming from our API users, are heavy and long. These one plus the smaller ones could have led to partial server errors due to servers overloaded.

That is to say, the API queries requested a specific setting process compared to the more classics (videos, web pages...). A few weeks ago, we have added load balancers servers in order to separate the API queries. 

If you are a Libcast user but you don't use the API: your browsing is slightly better. 

If you are a API user, I would highly recommend you to read the following:

  • Following some of our users requests, in case of error on one of our front-end servers, you will receive a XML error. 
  • The long queries (more than 20 sec), are from now on managed on the API servers. They are deliberately cut on the other servers to assure a better service. 
  • On the 15th November 2015, we will close the endpoint and the ones with your domain name (ex: You can change this endpoint to in your app. 
  • In case of growth in demand, we will be able to add servers easily, without any impact on your operations.

Access the Documentation

Follow the quality of our services


Furthermore, we have published

You can now evaluate in real time the quality of our services. To do so, we have implemented a color coding: 

  • Green : our service is fully operational. 
  • Orange : we have identified a slowdown on the platform. 
  • Rouge : the service is not available at the moment. 

Then, you can find the level of service on different categories: 

    • Frontends : refers to all that can impact your videos. (problems on the console, on code embeds etc.). 
    • Databases : refers to all that require the databases to work properly (permissions, statistics, etc.). 
    • Transcoding : you can evaluate here if the videos are quickly encoded. If this part is in orange, I would recommend you to look at the "System Metrics" underneath. It indicates in real time the number of videos waiting to be encoded before yours.
On this example, at 5.05 pm, 10 videos were waiting to be encoded

On this example, at 5.05 pm 10 videos were waiting to be encoded. 

    • Filesystem : refers to our storage capacity. 

Then, you can find the past incidents. If you experience a slowdown on the platform, I would highly recommend you to have a look at this section. Indeed, it is updated in real time.

Learn more

Do not hesitate to contact us for any technical request. 



CTO and co-founder