Implementing Open-source Observability within Maersk
summary
Maersk standardizes observability with a 100% self-service platform built on the LGTM stack (Grafana, Mimir, Loki, Tempo) plus in-house tooling and a read/write proxy (Guard) that enforces label rules and drops nonconforming data. Platform runs on ~11,000 CPUs and 40 TB RAM, ingesting ~120M active series, supports ~2,300 Grafana users, and enforces query protections to block monster queries. Plans include Thanos for multi-region querying and extending coverage to 700 vessels, 60 terminals, 450 warehouses while keeping 100% self-service.
AI-written summary · about
chapters (9)
- 0:00Maersk observability challenges
- 2:11Unified observability strategy
- 3:36Maersk Observability Platform
- 5:26Read-write proxy for data protection
- 8:17Self-service ingestion with Guard proxy
- 10:49Label-based rules for standards
- 13:14Monitoring dropped data
- 14:39Success metrics and expansion
- 15:45Edge observability for vessels
related talks
29mHow Maersk is Navigating the Seas of Observability with the LGTM Stack
28mMonitoring Stack was yesterday – Observability Platform at scale!
25mGrafana as Code-An Open Source Observability Solution
24mHow the LGTM Stack changed the observability culture at Wise Payments
23mImplementation of Unified Observability at Scale From Scratch
36mObservability at Scale With Monitoring as Code...
34mThe Grafana Stack: Building a Highly Available, Scalable Observability Platform
29mParanoid Observability –Success Stories and Misfortunes observing an observability stack