If the service availability exceeds the availability objective (little or no downtime) we have a positive Error Budget and we can continue to release more service updates.
Should the availability fall to the SLO, or lower, service releases must be put on hold. Recovering from a service failure ‘spends’ some or all of the Error Budget. The effect is that service development now have ‘skin in the game’ to produce a more reliable service or suffer not being responsive to their customer.
Launch and Iterate mantra
Aligned to agile development, agile reliability improvements should be delivered in regular, small, flexible packages. Have customer feedback sessions to confirm the enhancements are delivering what was planned.
Production Readiness Review
The equivalent of Operational Readiness checking. An
agreed structure of review to be applied to decide whether the quality and warranty aspects are acceptable. Three approaches presented, the strongest being the SRE Platform – a skeleton service structure (standardising reliability, security, availability, infrastructure, etc) with the service specifics being ‘plugged into’ this skeleton.
This reminds me of the ITIL v2 Application Management superstructure development, where service functionality is plug and play. Releases become small and localised, allowing flexibility and reduced risk.
Terminology mapping
“Problem Management investigation” is called “Post Mortem review”. The essential element is still a ‘No Blame’ culture.
“Capacity Management” is called “Service Intent Planning”.
Still, the mapping to ITIL practices is not a hard one.
Conclusion
I love the innovation of the new Site Reliability Engineering role. Rather than being concerned about the ever-changing IT world, Service Management professionals should embrace this new role as it aligns so well with existing (and future) ITIL thinking.
Bring it on!
About the Author
"Gary is an independent Service Management consultant and trainer. He has been a member of itSMF for over 12 years. Gary is a ITIL v3 Expert and v2 Master, with supporting knowledge in Agile, DevOps and Lean Thinking."
He would love to hear your feedback and if you would like to discuss the articles further please email Gary on: