DevOps / SRE · Cheatsheet

DevOps / SRE Cheatsheet

A short reminder sheet that pairs with the full playbook. Use it in a meeting when you need the next move, not a long read. Tap “What this means” on any card if a heading is unclear.

Civil Service specific Universal any sector Both

MAP-01

Reliability maps

Both

Flow: DORA · Continuous Delivery

Reliability: SLO · error budget · observability

Learn: incident command · blameless review

Secure: pipeline gates · supply chain

RL-01

The role

Both

Own the path to live and the truth of production.

Run daily:

Safe path to prod?
Know user pain early?
Is toil shrinking?

SL-01

SLOs & budgets

Universal

SLIs from user journeys
SLO agreed with product
Burn policy at 50/25/0
Review burn with roadmap trade-offs

DR-01

DORA & CD

Universal

Small batches · trunk · pipeline as path

Rollback rehearsed

Report DORA next to RAG

Do not game empty deploys

OB-01

Observability

Universal

Page on user impact + runbook

Delete noisy alerts weekly

Logs, metrics, traces with owners

Instrument in Definition of Done

IN-01

Incidents

Universal

Commander · tech · comms · scribe

Restore first, learn second

System actions with owners

Track repeat classes

TL-01

Toil & platform

Universal

Measure toil hours

Golden paths over ticket queues

Self-service with guardrails

Escape hatches are exceptions

SC-01

Secure pipeline

Both

High-signal SAST/SCA
Secrets detection + rotation
Signed artefacts, least privilege
Overrides timeboxed and logged

CM-01

Tone & ask

Universal

On-call: next action

SRO: impact, ETA, ask

Budget decision: proceed / slow / freeze

Drop "blip". State impact window.

QV-01

Operability bar

Both

Signals for new paths
Rollback path tested
Secrets and exposure reviewed
Runbooks when users feel change

GV-03

Troubleshooting: symptom to first move

Universal

!Pager storm: silence non-actionable; rewrite noisy alerts against SLOs

!War-room releases: shrink batch; rehearse rollback; kill manual default

!Budget ignored: write burn policy with PO/SRO; apply at next threshold

!One-person release: automate golden path; drill specialist absence

!Toil flood: cap interrupt; automate top toil source

!Repeat incidents: require owned system change before close

REF-01

Field reference

Both

One-minute checklist when production is on fire:

Name user impact, one sentence
Alerts, batches, SLO policy, toil or security gates?
Decision that restores or unblocks most, who owns it
Brief SRO / product before distorted versions
Protect on-call from noise while you sort it

Reliability: SLI, SLO, error budget, MTTR

Flow: DORA, CD, golden path

Learn: incident roles, blameless review, toil

Secure: SCA, SBOM, CAF, least privilege