Monitoring at Bedrock :
At Bedrock Streaming, a large part of our applications are hosted on Kubernetes clusters, others use the EC2 service from AWS and a small part are hosted on “OnPremise” servers.
Hi! We’re going to start our fourth article about Bedrock’s API gateway.
Today we will talk about the circuit breaker pattern, what it is, and how we’re using it.
Previously we discussed how we manage the load of our Kubernetes clusters and how we can anticipate our needs with prescaling. Today, we are here to share our solution that we have reworked and open sourced!
A quick sidetrack in our series about Bedrock’s API gateway.
This piece defines what are we talking about when we say “an error”, and explains how we handle the numerous connections to services we are calling.
C’est une histoire bien connue, dans la vie de n’importe quel développeur : un ticket arrive dans le backlog, décrivant une problématique relativement complexe. C’est parfois une question de technologie inconnue, ou parfois simplement un chantier un peu trapu. Je pense que toutes les équipes ont, au moins une fois dans leur vie, fait face à ce genre de tâche impossible : c’est l’occasion des regards désespérés, alors qu’un junior se lamente en disant « Mais par où est-ce qu’il faut commencer ? ». Et c’est là qu’on répond : « Essaye de faire un spike ».
Welcome to our second article about the backend architecture and its api gateway.
In the first part, we talked about the BFF and all services it depends on.
Today we’re going to take a look at what to do when one of them (or many), fails to respond.
At Bedrock, we have a CDN that outputs on average 400GB of uncompressed logs per hour. In this article, we present the architecture we have setup to collect these logs and extract value from them.
