SLA / SLO Error Budget Calculator
Calcola il downtime consentito e l'error budget per qualsiasi target SLO. Tabella completa dei "nines" di disponibilità.
Configura il tuo SLO
Esempi: 99.9 (3 nines), 99.99 (4 nines), 99.999 (5 nines)
Error Budget & Downtime
Downtime Consentito per Periodo
| Periodo | Downtime Consentito | In ore | In giorni |
|---|---|---|---|
| Giorno | 1.4 min | 0.024h | 0.0010gg |
| Settimana | 10.1 min | 0.168h | 0.0070gg |
| Mese (30.44gg) | 43.8 min | 0.731h | 0.0304gg |
| Trimestre | 2.19 ore | 2.190h | 0.0912gg |
| Anno | 8.77 ore | 8.766h | 0.3652gg |
Tabella dei "Nines" di Disponibilità
| Livello | Uptime % | Downtime Annuale |
|---|---|---|
| 1 Nine (90%) | 90% | 36.5 giorni/anno |
| 2 Nines (99%) | 99% | 3.65 giorni/anno |
| 2.5 Nines (99.5%) | 99.5% | 1.83 giorni/anno |
| 3 Nines (99.9%) | 99.9% | 8.77 ore/anno |
| 3.5 Nines (99.95%) | 99.95% | 4.38 ore/anno |
| 4 Nines (99.99%) | 99.99% | 52.6 min/anno |
| 4.5 Nines (99.995%) | 99.995% | 26.3 min/anno |
| 5 Nines (99.999%) | 99.999% | 5.26 min/anno |
SLA, SLO e SLI: le differenze
SLA (Service Level Agreement)
Contratto formale con il cliente che definisce le penali in caso di downtime. Il SLO interno dovrebbe essere piu stringente del SLA per avere margine.
SLO (Service Level Objective)
Obiettivo interno di uptime del team. Il target reale che l'ingegneria deve soddisfare. Tipicamente piu alto del SLA (es. SLA 99.9% → SLO 99.95%).
Error Budget
La quantita di downtime "consentita" prima di violare il SLO. Quando l'error budget e esaurito, si congela il rilascio di nuove feature e si focalizza sulla stabilita.
SLI (Service Level Indicator)
La metrica concreta misurata (es. % richieste con risposta < 200ms, % richieste 5xx). L'SLI e il dato; l'SLO e l'obiettivo sull'SLI.
Come utilizzare SLA / SLO Error Budget Calculator
Set Target Availability (SLO)
Enter the percentage of uptime you want to analyze (e.g. 99.9%) or choose a quick preset. The value must be between 0 and 100.
Read allowed downtime period
The calculator shows how much downtime is allowed per day, week, month, quarter, and year, while maintaining the same availability target.
Compare with the "nines" table;
Use the reference table to understand which standard availability classification (from 1 to 9) corresponds to your SLA.
Use error budget to manage releases
Error budget indicates how much downtime you can still "spend" in the current period. If it's exhausted, consider slowing down releases and focusing on system stability.
Suggerimenti
- Define your internal SLO more strictly than promised to the customer to have maneuvering room before a contract violation.
- Don't chase too many "nines": each additional availability digit exponentially increases the cost of infrastructure and redundancy needed.
- Monitor real-time error budget: enables swift response before exhaustion and balances release speed with stability.
Domande frequenti
What is the difference between SLA, SLO and SLI?
Service Level Indicator (SLI) is a concrete measurable metric, such as the percentage of requests with response under 200ms. Service Level Objective (SLO) is an internal engineering target for that SLI, such as 99.9%. Service Level Agreement (SLA) is the formal contract with the customer, typically having less stringent targets than SLO to allow for safety margins.
How is downtime allowed to start from the percentage of SLA?
Error budget is 100 minus the value of the SLO (e.g., with an SLO of 99.9%, error budget is 0.1%). Allowed downtime in minutes for a given period is calculated by multiplying the total duration of the period in minutes by the error budget percentage divided by 100. For example, with a month of approximately 43,800 minutes, an SLO of 99.9% allows for about 43.8 minutes of downtime.
What does "error budget" mean and how is it used in practice?
Error budget is the amount of downtime tolerated before violating the SLA within the reference period. As long as the error budget is not exhausted, the team can afford to release new features with controlled instability risk. When the error budget expires, SRE best practice advises freezing non-critical releases and focusing on system stability and reliability.
What does "3 nines" or "5 nines" availability mean?
Nines" refers to the informal way of referring to the number of 9s in uptime percentage: 3 nines correspond to 99.9% (about 8.77 hours of downtime per year), 4 nines to 99.99% (about 52.6 minutes per year), 5 nines to 99.999% (about 5.26 minutes per year). Each additional "nine" reduces the tolerated downtime by about an order of magnitude, and significantly increases the cost and complexity of necessary infrastructure.
Why should an internal SLO be more stringent than a contractual SLA?
If the internal SLO exactly matches the contractual SLA, any minor operational surprise (deploy issue, infrastructure downtime) would immediately trigger a contract violation with the customer. Having a more severe SLO (e.g., 99.9% SLA but 99.95% internal SLO) gives the team a safety margin to intervene before the problem becomes visible or contractually relevant.