There are very few things that truly require

[sujonkumar6300]

One of the key aspects of SRE is to avoid this altogether, by changing incentives to remove pressures between development and operations teams. Below are some of the strategies that Google uses to achieve this:

Have service agreements
Before a service can be supported, the level of availability it must maintain to keep users and the business happy must be determined: this is known as a Service Level Agreement (SLA). The SLA defines how availability is measured (e.g. percentage of queries successfully completed in less than 50 ms) and what the minimum acceptable value or SLO (service level sweden consumer email list objective value) is. This is a product decision, not a technical one. It is a very important number for an SRE team and its relationship with developers. It is not taken lightly, and must be rigorously enforced.

A 100% availability target is rarely a good idea, and in most cases is impossible to achieve. At companies like Google, a common availability target for online services is five nines (99.999%); this means that a service cannot be down for more than 5 minutes during a whole year.

100% availability (for example: a pacemaker), and achieving it is very expensive.

Measure and report performance
Once you have defined the SLA and SLO, it is very important that these are continuously monitored and reported. If you wait until the end of the quarter to produce a manual report, the SLA is of no use, since it is of no use to us to find out about problems so late.