How it works
How SteadyNines stops SLO breaches before they happen
SteadyNines runs a closed loop on every SLO: it detects error budget burn, finds out why, stops it with an approved action, removes the cause in code, and learns from the outcome. It starts on burn rate, not breach, so there is still budget to protect.
Why burn rate, not breach.
A breach alert tells you the promise is already broken. Burn rate tells you how fast you’re heading there. A 99.9% SLO over 30 days allows about 43 minutes of failure. At a burn rate of 14.4 that budget is gone in about two days; at 36, in 20 hours.
SteadyNines watches short and long windows together, so fast spikes and slow leaks are both caught early without paging on noise.
The five steps.
Detect.
SteadyNines reads SLIs from your existing telemetry and calculates burn rate for every SLO continuously. When a burn crosses your threshold, a situation opens.
Trace.
The Context Graph traces the burning SLO to the services, hosts, dependencies, deployments, config changes and open tickets behind it, and ranks likely causes with evidence.
Stop.
The AI SRE proposes the action most likely to stop the burn with the least risk and sends it to Teams or mobile with the evidence and the expected effect on burn rate.
Fix.
Once the burn is stopped, the AI Engineer opens the fix in the developer’s IDE: the suspected change, a drafted patch, tests and the full trail.
Record and learn.
Incident and problem records are written in your ITSM tool, and the situation is stored so the next matching one starts from what worked.
The same card-auth-api situation, minute by minute.
Burn detected at 02:14, stopped at 02:19, fixed in code by 09:12. Every step is timestamped and kept with the situation.
- 02:14burn 14.4× detected
- 02:15cause ranked: deploy #4821
- 02:16rollback proposed
- 02:18approved by on-call
- 02:19burn < 1×
- 02:40fix PR drafted
- 09:12fix merged
Questions
What does SteadyNines need to start?
Read access to your telemetry and SLO definitions, plus the tools you want it to act through. Discovery builds the Context Graph automatically.
Does it replace our alerting?
No. It sits on top of your monitoring and acts earlier than an alert would.
Can we import existing SLOs?
Yes. Talk to us about the SLO sources you use today, including OpenSLO definitions.