liveBook · Manning

Sidecar proxy

As explained in Grokking Software Architecture

A sidecar proxy is a supporting network proxy deployed alongside an application instance to handle communication and infrastructure concerns.

A sidecar proxy is an out-of-process network proxy that runs alongside an application container in a service mesh. It can take on network tasks without adding that logic to the application. section 10.4.2

Give the proxy work that can happen transparently, and keep business decisions with the application:

  • A proxy can retry dropped network packets, pool connections, use mTLS, and eject failing nodes from a load balancer pool. section 10.4.2
  • Use application code for domain-aware fallbacks, such as returning cached data or queuing an order; the proxy does not know the business Plan B. section 10.4.2
  • A service mesh can also use sidecars to handle mTLS and identity automatically, supporting Zero Trust in polyglot or legacy environments. section 11.10

How Grokking Software Architecture covers it

How deep it goes
Defines an out-of-process proxy beside an application container and names its network tasks, without showing proxy configuration.
Examples
Uses Envoy sidecars for transparent retries and contrasts them with application-owned business fallbacks.
What you can do after
Apply the service-mesh boundary by assigning transparent network tasks to proxies and domain-aware fallbacks to application code.
How it connects
Links service meshes with Envoy, Linkerd, Istio, mTLS, retries, connection pooling, circuit breaking, and Zero Trust.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Grokking Software Architecture. The section links point to the parts of the book it draws on.