infrHow infr thinks
The ideas behind the analysis, in plain words, and the calculators for the numbers behind them. Every figure is a real diagram run through the real engine, and every calculator uses the same formulas.
Start with single point of failure analysisQuorum calculator
Majority size, failures tolerated, and what a lost rack does.
5 nodes need3 upTolerates 2 failures
SLA and downtime calculator
Availability target to allowed downtime, in series and parallel.
99.9% allows8.77 hoursOf downtime a year
Quorum and failure domains
How quorum clusters lose their majority, how racks, zones and servers act as failure domains, and how infr finds a cluster that one failure can break.
Infrastructure availability
How infr turns component availability into a figure for your whole design: series and parallel, exact bounds over simultaneous failures, and when it simulates.
Error budget calculator
How much failure an SLO allows, and how fast it burns.
99.9% over 30 days allows43.2 minutesOf full outage
MTBF and MTTR calculator
Availability from time between failures and time to repair.
MTBF 1,000 h, MTTR 1 h gives99.9001%Availability
Rack diagrams
What can be mounted in a rack in infr, how rack space and power are checked, and what losing a rack does to everything inside it.
RAID and ZFS capacity calculator
Usable capacity, failures survived and rebuild for RAID and ZFS.
RAID 5, 6 drives of 4 TB gives20 TBSurvives 1 drive failure
Power and UPS runtime calculator
Rack draw against a budget, and UPS runtime at a load.
2,300 W at 230 V draws10 AAt maximum draw