· 1 min read
Postgres 14 leaves support in November
The five-year window closes on schedule. Two hundred words for whoever owns the managed database fleet.
#postgres#eol
Operations
10 articles
· 1 min read
Count the BI seats before the vendor does
License counts drift up; usage does not. The two-hour audit that funds itself, and the tier mistake hiding in most contracts.
#bi#licensing#cost
· 2 min read
Find the Databricks workspaces Microsoft upgraded to Premium for you
Azure retired the Standard tier on October 1 and moved every leftover workspace up a tier. How to find yours, price the change, and decide what to do with it.
#databricks#azure#cost
· 1 min read
Your data team's calendar is an architecture document
Read the recurring meetings the way you would read a systems diagram. The standing sync is a missing interface with attendees.
#operations#teams
· 1 min read
How to audit an analytics stack in two weeks
The method behind our assessment, published in full. Inventory consumers, trace lineage backwards, price every workload, interview the people on call.
#assessment#audit#operations
· 4 min read
How to staff a data platform without an org redesign
Central bottleneck or embedded sprawl is a false choice. The topology that works at 50 to 500 engineers, and the one hire that decides it.
#teams#operations
· 4 min read
Data SLOs that stakeholders actually understand
Freshness, completeness and accuracy promises, written in the consumer's units. How to define data service levels without importing SRE ceremony.
#slos#quality#operations
· 2 min read
Take the cost baseline before you change anything
Modernization without a baseline is a story, not a result. What to measure in the week before any platform work begins, and why finance will thank you.
#cost#assessment
· 4 min read
What happens inside a Data Platform Assessment
The two-to-three week engagement, day by day. What we need from you, what you get back, and the two outcomes we consider a success.
#assessment#engagement
· 2 min read
Executive dashboard anti-patterns we remove most often
The five patterns that make leadership dashboards untrusted, and the repairs. None of them require new tools; all of them require deleting something.
#bi#dashboards