Several posts this week touching on service discovery in distributed systems. It feels like with easier to use tools like Serf, and lower level systems like etcd we’re approaching a point where this will become, if not easy, then at least available to everyone. I’m pretty interested in how that changes the overall systems we design and build for do send me links about that if you come across any.
Devops Weekly is sponsored by Brightbox Cloud - launch cloud servers in multiple UK datacentres in seconds…
A excellent piece about team structures that support a devops approach. Good examples of why different structures work, and examples of others that don’t.
Knight Capital managed to lose $460m due to a software problem last year, and the administrative proceedings from the SEC litigation have been published. Anyone interested in how systems fail should read the full report, but the blog post highlights a few interesting parts.
Powershell has a number of nice features and with the release of Powershell 4 it has introduced Desired State Configuration (DSC). This works in a very similar way to Puppet or Chef, with the ability to regularly check a central source for configuration information.
Good slides explaining some of the advantages of a micro-services architecture, as well as discussing why functional programming works well in this context.
A good list of practices and problems to overcome when adopting devops in an organisation, focusing on some of the challenges found in a larger, more traditional, company.
Interesting use of DNS for service discovery. Discusses implementing a custom DNS server which uses SRV records to announce services and a simple JSON post to the DNS server to announce availability and update TTL on the DNS record.
A blog post about evolving a deployment and code merging strategy, specifically using CircleCI. Includes code examples and a good explanation of how the team incorporate deployment in their workflow.
FreeAgent are looking for a talented Operations Engineer to work on the FreeAgent app, internal services and supporting infrastructure. You’ll be working alongside our (remote) Ops team squashing single points of failure, fixing bottlenecks, profiling load and solving interesting scaling and automation problems.
Serf is described as decentralized solution for service discovery and orchestration that is lightweight, highly available, and fault tolerant. Lots of use cases from datastore clustering, load balancing web servers or triggering events across a dynamic cluster.
confd is a lightweight configuration management tool focused on keeping local configuration files up-to-date by polling etcd and processing template resources and reloading applications to pick up new config file changes.
The latest release of Collectd now supports collecting metrics via the StatsD protocol. If you’re running both statsd and collectd you may be able to run only the latter.