Links

• When to shard mongodb: A short list of reasons you may want to consider sharding. Sharding adds a bunch of complexity to a db cluster so deferring this as long as you can is often a great choice. (Many datasets never get big enough to warrant the extra complexity.) Things like disaster recovery time to restore service, available resources on a single machine, hot data + cold data and a few more. Pretty nice list here!
• Road to sre: We have a new acronym! ROAD -> response, observability, availability, delivery. The author is describing the responsibilities of the sre team at his company. Sounds like they participate in the incident response team (including with metric tracking for frequency, severity and duration of events) as well as defining slis and slos and delivery pipelines! (Build pipelines, infrastructure as code, deployment tooling which has been fairly typical in terms of responsibilities for an ops team at the companies I’ve worked.)