Step 2: Metadata
Pick a a clear and explicit name for your SLO. It is important to choose appropriate names and descriptions that effectively communicate the goals and expectations of the SLO. Well-crafted SLO names and descriptions provide vital information and are easy to understand and communicate.
A good rule of thumb for naming an availability SLO is to use the following template:
{target value}% of requests to {service operation} are processed successfully
This template emphasizes the desired success rate, the service operation, and the primary goal of ensuring successful request processing. By including the target value, the SLO name provides a concise overview of the availability expectation.
Similarly, for latency SLOs, a recommended naming template is:
{target value}% of successfull requests to {service operation} are processed under {threshold + units}
This template focuses on both the success rate and the maximum acceptable latency for requests. By including the threshold and units, it provides additional information about the expected response time. This helps set appropriate expectations and allows stakeholders to understand the trade-off between performance and success rate.
Categorize your SLO in terms of the performance area that your SLO is measuring. If you selected an SLI recommendation, a category will already be selected by default. The supported categories are as follows:
- Availability
- Latency
- Quality
- Correctness
- Freshness
- Throughput
- Durability
Pick the member of your team that should own the SLO. This is the person to reach out to know more about the SLO and regarding accountability.
The owner is set to the current user by default.
Custom labels are key-value pairs that you can use to provide contextual information to the SLO. They will be used across the application for filtering and grouping purposes. By attaching labels to SLOs, you can easily categorize and organize SLOs based on specific dimensions that are of importance to your organization, allowing for more effective monitoring and analysis.
Last modified 28d ago