An error budget turns 'is it reliable enough' into a number
99.9 percent over 30 days allows 43 minutes of failure. Spend it on incidents and there's none left for risky deploys. Spend none of it and you're being more careful than your users asked for.
Graph how fast the budget is burning. When it's running out, the team ships reliability work instead of features, because you agreed on that in advance and not in an argument in the middle of an incident.
observabilitysre