// sla sources
The dataset.
Every value has a source and a lastVerified date, and CI fails when one goes stale.
Wrong or outdated? Send a PR — it's one YAML file.
standard
99.900%
8.8 h/yrImportant: the SLA requires >=2 replicas for reads and >=3 for writes. A single-replica service has NO SLA — check your replica count.
standard
99.900%
8.8 h/yrProvisioned throughput deployments may carry different guarantees; multi-region routing is the usual resilience upgrade.
zone-redundant
99.990%
0.9 h/yrRequires Premium v3/v4 and at least 2 instances.
single-instance
99.950%
4.4 h/yr
v2
99.950%
4.4 h/yrAssumes >=2 instances or autoscaling enabled.
standard
99.900%
8.8 h/yr
zone-redundant
99.950%
4.4 h/yr
standard
99.900%
8.8 h/yr
standard
99.950%
4.4 h/yrZone redundancy is configured on the managed environment; verify env-level ZR for zone resilience.
standard
99.950%
4.4 h/yrCovered by the Container Apps SLA — verify jobs are included in the current SLA wording.
standard
99.900%
8.8 h/yrPremium tier supports zone redundancy and geo-replication.
standard
99.900%
8.8 h/yr
zone-redundant-ha
99.990%
0.9 h/yr
same-zone-ha
99.950%
4.4 h/yr
no-ha
99.900%
8.8 h/yrHA requires General Purpose or Business Critical tier; compute cost roughly doubles.
zone-redundant-ha
99.990%
0.9 h/yr
same-zone-ha
99.950%
4.4 h/yr
no-ha
99.900%
8.8 h/yrHA requires General Purpose or Memory Optimized tier (not Burstable); compute cost roughly doubles.
standard
99.990%
0.9 h/yr
standard
99.990%
0.9 h/yr
standard-with-zones
99.950%
4.4 h/yr
standard
99.900%
8.8 h/yr
free-tier
99.500%
43.8 h/yrFree tier is an SLO, not a financially-backed SLA. Standard tier adds the uptime SLA.
standard
99.990%
0.9 h/yr
premium
99.950%
4.4 h/yr
standard
99.900%
8.8 h/yr
zone-redundant
99.995%
0.4 h/yr
standard
99.990%
0.9 h/yr
zrs
99.990%
0.9 h/yr
geo-redundant
99.990%
0.9 h/yr
lrs
99.900%
8.8 h/yrLRS lives in a single datacenter. ZRS spreads across zones at moderate extra cost.
zonal-single
99.900%
8.8 h/yrSingle VM in a zone. Add a second VM in another zone + LB and mark as redundant set (99.99% as a set).
single-premium-disk
99.900%
8.8 h/yrAssumes Premium SSD / Ultra disk. Standard HDD single VM carries a lower SLA — verify.
Deliberately not counted
known types, no own SLAmicrosoft.web/sites
Availability is determined by the App Service plan (microsoft.web/serverfarms), which we score. Counting both would double-count the same dependency.
microsoft.insights/webtests
Monitoring configuration, not a runtime dependency of your application's availability.
microsoft.app/agents
Preview service without a published SLA (as of last verification). We don't guess.
microsoft.app/managedenvironments
SLA applies to the container apps inside it, not to the environment resource itself. Its zone redundancy setting affects the apps' resilience.
microsoft.insights/metricalerts
Management-plane configuration, not a runtime dependency of your application's availability.
microsoft.network/networkwatchers
Diagnostics tooling auto-created per region — not a runtime dependency of your application's availability.
microsoft.sql/servers
The logical server is a free management container — the SLA applies to the databases inside it, which we score individually.