Definition:
MTBF — Mean Time Between Failures The average time a system runs correctly before its next failure. It measures reliability. A higher MTBF means the system breaks less often.
MTTR — Mean Time To Repair / Recover The average time needed to diagnose a failure and restore full service. It measures resilience. A lower MTTR means the system heals faster.
Example:
The Coffee Shop Espresso Machine You run a café and track your primary espresso machine for one year.
Incident Date Downtime
Pump seal blows Feb 1 4 hours
Grinder motor fails Jun 1 2 hours
Heating element cracks Nov 1 2 hours
MTBF (How reliable is it?)
The intervals between failures are ~120 days and ~150 days.
MTBF=
2
120+150
≈135 days
Meaning: You can count on roughly 4.5 months of uninterrupted service between breakdowns.
MTTR (How fast do we recover?)
The repair durations are 4, 2, and 2 hours.
MTTR=
3
4+2+2
≈2.67 hours
Meaning: When the machine does die, your technician gets it running again in under 3 hours on average.
The Key Takeaway
Question Metric Goal
How often do things break? MTBF Higher (fewer incidents)
How fast do we fix them? MTTR Lower (faster recovery)
In software, teams often focus only on preventing crashes (MTBF), but world-class SRE teams also drive MTTR down through automated rollbacks, feature flags, and redundancy—because every system eventually fails, and what matters is how little your users notice.