Toil Identification & Reduction

What is “Toil”?

In Site Reliability Engineering (SRE), toil is the kind of work tied to running a production service that tends to be:

• Manual: Typing terminal commands, resetting passwords, executing repetitive runbook steps.
• Repetitive: Tasks performed repeatedly rather than one-time design work.
• Automatable: Tasks that could be performed by a computer script or automated workflow.
• Tactical: Work reactive to immediate alerts rather than strategic long-term architecture.
• Devoid of Enduring Value: If you complete the task today, the system is in the exact same state tomorrow without permanent improvement.
• O(N) with Service Growth: Tasks that scale linearly with user traffic or server count.

What is NOT Toil?

• Administrative overhead (team meetings, emails, HR training).
• Designing new features or writing automation code.
• Resolving novel complex incidents requiring architectural investigation.

The SRE 50% Rule

Google mandates an upper bound of 50% of an SRE’s time on operational toil (on-call shifts, tickets, manual tasks).

The remaining $\ge 50%$ must be protected for engineering project work:

• Writing software to automate away common failure modes.
• Building resilient self-healing infrastructure.
• Implementing telemetry and observability.