Rely.io
Search…
⌃K

Create an SLO

Learn about the different ways you can create the SLOs you want.
To start creating an SLO you first need to open the SLO Wizard. You can do this by clicking the Create SLO button at the top right corner of the SLO Workbench.
Accessing the SLO Wizard

Step 1: Define your SLI

As the first step of the SLO configuration process, you start by defining your SLI. The SLO Wizard allows you to do this in three distinct ways:
  • SLI Recommendations: select an out-of-the-box SLI generated by Rely.io's recommendation engine.
  • Metrics Explorer: explore and transform the metrics you made available through your data source integrations to create an SLI from scratch
  • Query: directly query your data sources for custom metrics in their native query language.
The following documentation goes through each one of these methods in detail.
SLO Wizard first step

SLI Recommendations

You can always start by looking through Rely.io's SLI recommendations list to see if any of them suit your SLO use case. Recommendations are grouped by category which informs you about what performance area the SLI in that category measuring. Within each category group, you can find the actual SLI recommendations. The recommendation's name tells you what the SLI is measuring while the labels next to the name identify the underlying monitoring metrics used in its calculation.
Click on one of the SLI recommendations to select it. Doing so, should trigger the SLI formula to appear with the metric fields already pre-populated with the monitoring metrics being used to caculate the SLI.

Metrics Explorer

If you weren't able to find a suitable SLI amongst the various recommendations provided by Rely.io you can try a more hands-on approach to building and creating your SLI with the Metrics Explorer. The Metrics Explorer gives you access to all the monitoring metrics Rely.io is ingesting through your data source integrations. You can apply different filters to this list in order to declutter what can sometimes be a daunting list of metrics to go through. Filter metrics according to their dimensions so that you can focus your attention on the data that actually fits the purpose of the SLI you intend to create.
The Metrics Explorer also allows you to manipulate existing time series so that you can build the metrics you whish to use in your SLI calculation. To do this, follow the steps bellow for each metric you are looking for:
  1. 1.
    Select a metric name: the Metric Name selector displays all the names for the time series made available by the data sources you have selected.
  2. 2.
    Group multiple time series into sets that match selected labels: use the Group By input to create a time series for each combination of matching labels.
  3. 3.
    Optionally, experiment with different aggregation functions: chose from a list of available aggregation functions how you want aggregate the data points of the time series' groups you created in the previous step.
After you're satisfied with the metrics you have built, you can then select one, or more to add to your SLI formula. Selecting a metric automatically displays it in the Metric Preview chart. You can then repeat this process to build any other metrics you want to use on your SLI.
Query Selector
Finally, the SLO Wizard offers you a third way to find the metrics for your SLI: direct queries. If you have any data source integrations that support direct queries, you can select this option to create your own custom query for retrieving the data that you're looking for. After you write down your query, click the Fetch metric and preview button at the bottom of the query input. This will trigger Rely.io to run your query and retrieve the metric associated to it. If the metric has any historical time series data, it will be displayed in the Metric Preview chart automatically. You can then add this metric to the SLI formula.
This functionality is currently only enabled for New Relic data sources with some restrictions with regards to the clauses that can be used in the querys.
SLI Formula
You can chose between two formats for your SLI formula, either Ratio, or Threshold.
  • Ratio: defines an SLI as the proportion of good events compared to the total number of valid events and it’s a better fit for count-based metrics. When you select this format, you need to specify the metrics, or combination of metrics that you consider to be Good Events and Valid Events so that they are then used for the SLI calculation.
  • Threshold: defines an SLI as the proportion of calls that are above, or bellow a certain threshold value against the total number of calls and it’s a better fit for time-based metrics. When you select this format, you need to chose the metric you want to track and then specify a threshold value for that metric. You can also edit the threshold’s comparison operator by selecting on of two options: > (greater than), or < (lower than).
Every time you select a Threshold SLI, you are prompted to specify a threshold value. To help you do this, you can leverage the Metric Preview chart by clicking on the Preview Metrics button in the top right corner of the SLI formula element. This displays the metrics you have assigned to your SLI and allows you to visualize how these metrics have been evolving over. You can then leverage this information to help you find an appropriate threshold value for your SLI.
The SLI shown in the SLI Preview chart is always calculated using a 7 day rolling window. Take this into consideration when interpreting this chart.
Click here to learn more about how rolling windows work at Rely.io.

Selecting data sources

Regardless of the method you chose, you can select wich data sources you want to retrieve data from by using the Data Sources selector, displayed bellow the preview chart in every panel. This allows you to explore different metrics in the same place and build SLIs based on your various data source integrations.

Step 2: Compliance Window and Target

After defining your SLI, you can move on to the next step in the configuration process. In this step, you will define your service level objective. To do this, you must specify two parameters:
  • Compliance Window: this is a default rolling time window for which your SLO will be continuosly evaluated against its target.
  • Target: determines what portion of valid events should be good events for Ratio SLIs, or what portion of time your SLI should be exhibiting good behaviour, for Threshold SLIs.
SLO Wizard second step
To configure your SLO's rolling window you must specify a single duration parameter that determines the size of the window. This window rolls at the end of each day. Then, specify the SLO target. This is a percentual number for which your SLI will be evaluated against for the duration of the compliance window.

Step 3: Name, Dimensions and Category

In the last step of the SLO configuration process you name and provide some context to your SLO. Start by specifying a clear and explicit name for your SLO. You can then assign your SLO to multiple dimensions. Dimensions are meant to provide context to your SLO within your application. We recommend you specify at least two dimensions: Service and User Journey. This would tie your SLO to a specific user-facing dimension (User Journey) and to your infrastructure (Service) so that you can correlate the perforamnce of the two. Learning how to correctly assign dimensions to your SLOs is critical since that will allow you to utilize your SLOs to look at your application's reliability from different prespectives.
Lastly, categorize your SLO. This category determines the performance area that your SLO is measuring. Some of the most commnon performance areas are Availability and Latency. If you selected an SLI recommendation, a category will already be selected by default, but you are still free to change it.
SLO Wizard third step
Finally, click the Create SLO button at the top right corner of the page. If everything was configured correctly, you should be able to see your newly created SLO in the SLO Workbench. You should also be able to immediately see how your SLO performed for the last compliance period since Rely.io computes the SLO's historical performance after it is created.