Service Discovery

Consul is a service discovery technology that ensures microservices can communicate with each other.

Distinguishing features #

Consul has some features that set it apart from other service discovery solutions.

HTTP REST API & DNS support #

Consul has an HTTP REST API and supports DNS.

• DNS (Domain Name System) is the system that maps host names such as www.innoq.com to IP addresses on the Internet.
• In addition to returning IP addresses, it can return ports at which a service is available.
• This is a feature of the SRV DNS records.

Configuration file generation #

With Consul Template, Consul can generate configuration files.

• The files may contain IP addresses and ports of services registered in Consul.
• Consul Template also provides Consul’s service discovery to systems that cannot access Consul via the API.
• The systems have to use some kind of configuration file, which they often already do anyway.

Health checks #

Consul can perform health checks and exclude services from service discovery when the health check fails.

• For example, a health check can be a request to a specific HTTP resource to determine whether the service can still process requests.
• A service may still be able to accept HTTP requests, but it may not be able to process them properly due to a database failure.
• The service can signal this through the health check.

Replication #

Consul supports replication and can ensure high availability.

If a Consul server fails, other servers with replicated data take over and compensate for the failed server.

Multiple data centers #

Consul also supports multiple data centers.

• Data can be replicated between data centers to further increase availability and protect Consul against the failure of a data center.
• The search for services can be limited to the same data center.
• Services in the same data center usually deliver higher performance.

Service configuration #

Finally, Consul can be used not only for service discovery, but for the configuration of services.

• Configuration has different requirements.

• Availability is important for service discovery.

• Faulty information can be tolerated.

• If the service is accessed and is not available, it doesn’t matter. You can simply use another instance of the service.

• However, if services are configured incorrectly, this can lead to errors.

• Correct information is more important for configuration than for service discovery.