Don’t Let Observability Inflate Your Cloud Costs

A shift occurred in the technology sector this year, as companies focused on the cost perspective of sustainability in the infrastructure and tooling space. Observability tooling, perceived as a large cost center, came under scrutiny, leading to the recommendation of allocating 20-30% of total infrastructure spend to observability. Tweaks in trace locality, such as keeping requests within a single availability zone, and employing sampling techniques like head and tail sampling, can help reduce costs in the observability stack. Additionally, considering peering and transit gateways, as well as utilizing vendor support for services like AWS PrivateLink or Azure private endpoints, can further alleviate cloud costs.