LATEST Articles
The real cost of blind deploys
We looked at deploy-related incidents across 200 teams. The time between a change shipping and an engineer knowing what broke averaged 47 minutes. That window is the problem.
Read ArticleWhat shipped in STRAND v2
Change impact analysis is live. You can now see everything a change will touch before it ships. We also redesigned onboarding from scratch: first mapped service in under five minutes.
Read ArticleArkos cut incident response time by 60%
Arkos was running five monitoring tools and still losing 15 minutes at the start of every incident just getting oriented. Here is what changed.
Read ArticleBuilding a dependency graph that stays accurate
Mapping services once is straightforward. Keeping the graph accurate six months later, after 40 teams have shipped changes, is a different problem entirely. Here is how we solved it.
Read ArticleYour MTTR number is probably wrong
What gets measured as recovery time usually includes 20 to 30 minutes of investigation before any actual fixing starts. That is not recovery. Separating the two changes what you optimize for..
Read ArticleFour tools and still flying blind
Four monitoring tools, none of them aware of each other. Every incident starts with the same manual process of reconciling what each one is showing. We looked at why this keeps happening.
Read Article